key: cord-0059547-lur7w11v
authors: Shafer, Byron E.; Sawyer, Elizabeth M.
title: The Usual Suspects: Strategies and Sequences
date: 2020-09-03
journal: Eternal Bandwagon
DOI: 10.1007/978-3-030-51799-1_6
sha: ad4e74456e6ebb3c5a611eeeecbd5dee523e80af
doc_id: 59547
cord_uid: lur7w11v

The other main purported influence on the politics of presidential nomination—preferred by pundits and professionals over bandwagons, seedbeds, or factions—are the strategies of presidential aspirants and the sequencing of state contests. So this chapter turns to these potential alternative influences, focusing first on the strategic actions of presidential contenders, then on the strategic actions of state political parties. Candidate strategies are more numerous, more varied, and more nimble, while state strategies are more grand, more slow-moving, but with more automatic impact. Yet again, they fail to derail an underlying bandwagon, to alter the occupational seedbeds of the contenders, or to neutralize the factional struggles that ultimately produce two individual nominees.

participatory caucuses. These purportedly influential factors, the ones that do attract some public attention, are the strategies that reliably characterize the nominating campaigns of presidential aspirants and that, at one remove, characterize the efforts by state political parties at augmenting their own nominating influence.

Like rules of the game, these strategic factors have the potential to alter one or another given outcome on the way to a presidential nomination. At a minimum and very concretely, the combination of this second set of maneuvers, that is, the interaction of candidate strategies and state calculations, is what produces the specific sequence of contestsand eventual outcomes-on the way to a presidential nomination. In this inter-action, state parties seek to optimize the influence of their delegate apportionment, just as presidential contenders seek to maximize their own allocation of these delegates. So state calculations do recurrently shape the sequence of contests within which the nominating bandwagon must role, while the strategies of contenders who respond to this sequence inevitably provide the detailed politicking that characterizes any given presidential year.

From one side, the sequence by which the key events of a nominating contest unfold will powerfully shape the strategies-mutual and interactive, after all-which the various participants produce. From the other side, the strategies of a field of presidential contenders will always alter the practical implications of any particular sequence of events. So the most concrete and specific way to organize a search for the contributions of both is to focus first on the actions of presidential contenders, then on the actions of state political parties. Candidate strategies are more numerous, more varied, and more nimble, in the sense of changing more quickly. State strategies are larger and more slow-moving, but exhibit a greater (and continuing) impact when they do occur.

These two moving pieces, being the ones highlighted by commentators of all sorts, are the other main purported influences on the politics of presidential nomination, even if they usually turn out to be a minor element in any comprehensive analysis. In the eyes of all too many observers, they appear to determine the final choice between one specific individual and another. And they may very occasionally be the final advantage or disadvantage for a specific aspirant in a particular year-in effect the remaining influence, the error term if you will, that necessarily completes a comprehensive analysis. Regardless, for an observer who does not care about how nominating politics works but cares a great deal about who turns up at the end of the causal funnel-not to mention for the individual aspirants themselves-these lesser factors may be the main point of the entire exercise. So an analysis that focuses on the bandwagon dynamic, candidate seedbeds, and structural factors need to pay attention to these other purported influences too.

Every potential candidate must decide whether to enter the nominating contest, how to enter the nominating contest, and when to enter the nominating contest. Because candidates vary widely in the strategic answers that are practically available to them, but also because the three answers are usually linked, observers tend to register and remember only whether any given individual was "in" or "out," along with their eventual fortunes. Yet while strategies do inevitably change as the contest unfolds, initial strategies ordinarily contain answers to all three opening questions. These almost inevitably emerge out of occupational backgrounds and prior careers. They are colored by the key structural influences shaping candidate selection. And they follow from judgments about the potential response of state parties, as well as of the other major players, who are simultaneously making judgments on the same matters.

The universe of those who seriously considered, or might well have considered, or should have considered the run for a presidential nomination remains infinite and unknowable. Moreover, the most basic strategic decisions made by aspiring candidates for a presidential nomination are perhaps the ones most quickly lost to history. Beyond this historical amnesia, most election years produce some actual entrants whose calculations could neither have been anticipated nor perhaps even comprehended. Just as many years feature the absence of one or two candidates who might have been expected to enter and were widely judged to be capable of doing so. Yet as diverse and amorphous as this process may seem, specific answers can be given a consistent patterning across time by the three major institutional regimes for presidential nominating politics: the pure convention system, 1832-1908; the mixed system, adding presidential primaries to state conventions, 1912-1968; and the modern plebiscitary system, dominated by explicitly candidate-based primaries, 1972-2020. Perhaps the outstanding modern example of the decision not to run, as well as a good introduction to the analytic puzzle and its practical implications, involved Mario Cuomo, Democratic Governor of New York from 1983 to 1994. Widely viewed as the front-runner for the Democratic presidential nomination of 1992, with substantial support from both party activists and the regular party, Cuomo did not enter. 1 So that contest ultimately produced the far less anticipated nomination of Bill Clinton, Governor of Arkansas, accompanied by frequent assertions in its aftermath that the absence of Cuomo was a major factor in the success of Clinton. Though the "ghost" of Mario Cuomo was even longer-lived than this headline example, running from his keynote address at the 1984 convention, to widespread elite speculation in 1988, to his nominating speech for Clinton in 1992, through his frequently referenced attack on congressional Republicans at the 1996 gathering.

An earlier example, with actual and more evident measures of the real impact of a nonentry, involved Samuel J. Tilden, reform Governor of New York and Democratic nominee for president in the disputed general election of 1876. 2 Having been deprived of what many regarded as a victory in that election, Tilden was widely expected to make a second run in 1880, and informed opinion dubbed him the incipient front-runner. Yet Tilden kept his intentions to himself, to the extent that when the 1880 Democratic convention convened, supporters had to settle initially for a "stalking horse," Henry J. Payne, former congressman from Ohio. Payne was expected to play that role only until Tilden entered. Instead, the role itself disappeared with Tilden's public statement that he would not seek the nomination, which in turn allowed Gen. Winfield Scott Hancock, the delegate front-runner at that point, to proceed to a nomination.

The modern plebiscitary era has converted choices by non-entrants into less of a topic for analysis. Because candidates must enter in order to acquire delegates, there are no longer any early non-entrants who are merely waiting for their moment. Instead, having demurred, non-entrants quickly become irrelevant. Delegates who have been acquired as personal loyalists by one candidate can rarely be acquired by another, while the bandwagon will begin rolling early among those who did not hesitate to enter. Moreover, this compression of the timeline for entry appears to have achieved a second level of impact: the modern era is much harder on what were once the "perennial possibilities," that is, major figures who might enter in one or more of an extended sequence of quadrennial contests.

Both the pure convention period and the mixed system still generated such figures, who seemed a perennial possibility for a presidential nomination, to the point where other aspirants had to take their expected actions into account. In that sense, they were omnipresent in collective strategic decisions and not just on their own. In the years before the Civil War, Henry Clay of Kentucky, Speaker of the House, became the prime example. Clay stood first as the presidential candidate of the National Republicans in 1832; endorsed the three-candidate strategy of the Whigs in 1836 and otherwise stayed clear of the presidential race; actively sought the Whig nomination in 1840 but was defeated, in part by imposition of the unit rule; was unanimously chosen as the Whig nominee in 1844; and came into the 1848 convention as the front-runner, before the supporters of Gen. Zachary Taylor and Gen. Winfield Scott united to deny him the nomination. 3 In the years after the Civil War, this role was instead played by James G. Blaine of Maine, likewise Speaker of the House. Blaine brought a wide lead into the Republican convention of 1876, before opponents united behind Rutherford B. Hayes as a dark horse. He was the lead opponent of a third nomination for Ulysses S. Grant in 1880, before the ultimate strategy for avoiding Grant came to involve uniting behind James A. Garfield. Blaine became the nominee in 1884, after the death of Garfield and over the opposition of sitting President Chester Arthur. He took himself out of the running in 1888, thereby encouraging multiple candidacies and producing the unexpected nomination of Benjamin Harrison. And Blaine seriously considered challenging the sitting but unpopular President Harrison in 1892, before deciding that this would guarantee electoral defeat for either of them. 4 For the late nineteenth century and comfortably into the twentieth, the potential contender who always had to be considered was William Jennings Bryan, former one-term congressman from Nebraska. Bryan stormed to an unexpected and highly contested Democratic nomination in 1896; was renominated without opposition in 1900; announced his noncandidacy in 1904, then tried unsuccessfully to block the eventual choice, Judge Alton B. Parker; retook control of the convention and its nomination in 1908 over only token opposition; did not stand in 1912 but threw his support to Woodrow Wilson, the eventual nominee; lost a contest for a delegate seat at the 1916 convention, which invited him to speak anyway; and returned as a delegate but otherwise as an operational shadow of his former self in 1920 and 1924, proposing various major alternative platform planks that were variously ignored or crushed. 5 No one after 1920 would have the strategic omnipresence of a Bryan, a Blaine, or a Clay, but the general phenomenon continued through the mixed system of presidential nominations. Thus in the 1940s and into the 1950s, each major faction of the Republican Party generated a continuing-a repeat-champion. For the moderate regulars, this was Governor Thomas E. Dewey of New York; for the ideological activists, it was Senator Robert A. Taft of Ohio; and the struggle to elevate or to displace one or the other would color Republican nominating politics for a long time. 6 Dewey, then Manhattan District Attorney, and Taft, already a Senator, inaugurated this struggle as front-runners at the 1940 convention, before insurgent Wendell Willkie elbowed first Dewey and then Taft aside.

Dewey, now Governor of New York, came back as the successful consensus choice in 1944, successful, that is, in the nomination but not the election. Dewey sought the nomination for a third time in 1948, again challenged by Taft, whose inability to forge a coalition with the other candidates, especially former Governor Harold E. Stassen of Minnesota, ultimately presented Dewey with his second successful nomination. And in a close contest in 1952, Taft returned as the champion of the activist wing of his party to stand against Dwight Eisenhower, who had replaced Dewey as champion of the regulars, while Stassen returned to lead the shift to Eisenhower, which completed the nomination of the latter. That did prove to be an end of a mini-era: thereafter, no candidate would seriously enter the strategic calculations of other possible contenders for four or more consecutive conventions, at least as this is written.

Conversely, if those men produced the most unpredictable contributions to candidate decisions in nominating politics-other potential contenders could know that these repeat entrants were potentially disruptive; they just could not know what they would do in any given year-then the least conditional contributions came from sitting presidents. Across time, most sought renomination, meaning that all other potential contenders had to begin with an estimate of the likely fortunes of the incumbent President (Table 6 .1). Two sitting presidents did actually commit to a single term as part of their initial nomination, James K. Polk in 1844 and Rutherford B. Hayes in 1876; both honored that commitment. One other, Andrew Johnson in 1868, ultimately sought the nomination of the other party, albeit unsuccessfully. Everyone else sought renomination and most achieved it, though it was the pure convention era, not the plebiscitary era, which produced the exceptions this time: 8 But four years later, he had acquired a record on the same poisonous disputes over slavery and statehood that had doomed Fillmore, and he too could not reconstruct a nominating majority from a factionalized party that needed desperately to avoid those poisonous issues. • Another unity ticket, for the nascent Republicans this time, appeared in 1864, when Abraham Lincoln as the Republican nominee for president took Andrew Johnson, a dissident Democrat, as his vice president. Lincoln was to die in office; Johnson was to ascend to the presidency; and his dissent on Reconstruction would put him so clearly at odds with his nominal party that Johnson actually stood for renomination at the Democratic convention in 1868 before his support ebbed away. • The only other sitting president to try and fail at renomination was Chester A. Arthur, vice presidential nominee in 1880 on a Republican ticket with President James A. Garfield, but once again undone by the curse of being a ticket-balancer. Arthur too ascended to the presidency on the death of Garfield, but in 1884 much of the party preferred James G. Blaine, leaving Arthur-too regular for the reformers, too reformist for the regulars-with a substantial base in the skeletal southern Republican delegations but without the possibility of renomination. • The other incumbent presidents who ascended to the presidency on the death of the president-Theodore Roosevelt in 1904, Calvin Coolidge in 1924, Harry Truman in 1948, and Lyndon Johnson in 1964, along with Gerald Ford, who ascended on the resignation of Richard Nixon-would all content themselves with completing the terms of their predecessors and then seeking a further full term of their own. In that sense, they were never denied renomination. All but Ford secured that further term, though the prospects of Truman and Johnson would have been cloudy had they sought a second full term. 9 Said differently, after the pure convention era, incumbent presidents were always renominated if they sought to be so. The demands of party formation and the stresses associated with what became the Civil War did produce multiple exceptions in that first institutional period. Though even these always coincided with the death of the president who had actually been elected. No elected incumbent standing for renomination was denied that honor even during this first formative era. In any case, even these exceptions would disappear by the successor period, the mixed system of presidential nominations, and they would stay gone in the modern period, the plebiscitary era. Those are the messages of Table 6 .1.

Successfully renominated presidents did not always go on to win the general election. Martin Van Buren in 1840, Grover Cleveland in 1888, Herbert Hoover in 1932, and Jimmy Carter in 1980 all failed in November after succeeding in their respective conventions. Yet more to the point here, bracketing the renominations of sitting presidents highlights a much more varied story of presidential nominating struggles and results among the much larger cohort of nominees who remain. Though again, these individual stories too present some common overall patterns when further arrayed according to the three great institutional periods of presidential nomination. So Table 6 .2 goes on to underline the very real differences by era when there was no sitting president. Seen this way, it becomes clear that in the two earlier periods, containing the bulk of American political history, the majority of presidential nominations were not confirmed on the opening ballot. They required successive ballots, sometimes realizing the purported lead of the apparent front-runner on the opening ballot, sometimes elevating an entirely different candidate who at least had serious support when the convention opened, and sometimes finding an ultimate nominee who was not even in evidence when the opening ballot took place. Only the modern era would see all nominations without an incumbent president effectively realized in advance of the convention, that is, within the process of delegate selection. In fact, no convention in either party during this period was to require so much as a second ballot to confirm the result of that prior process.

Yet there were further strategic calculations specific to each period. In the pure convention era, the further question for a strategic presidential aspirant was whether to attempt to be a serious presence on the opening ballot or to adopt a strategy of being "available" but not demonstrably active. Under the mixed system, not otherwise very distinctive from the first era in terms of actual outcomes, an additional strategic question arose-whether to enter the presidential primaries-and this became a further calculation piled on top of the continuing choice about a general level of activity in seeking the nomination. By contrast in the plebiscitary period, where waiting to be recognized for your "availability" disappeared as an option and where mandatory primaries were effectively ubiquitous, the sole remaining strategic choice boiled down to entering one, the other, or both of two opening contests

In the pure convention period, the first three candidates who were not even put in nomination on the opening ballot but ultimately ended up as the nominee were all Democrats, though none of them featured much in the way of a calculated strategy:

• James K. Polk, the inaugural dark horse in 1844, did not appear until the eighth ballot but was nominated on the ninth. A former state official in Tennessee, his nomination was largely a response to the continuing deadlock between Martin Van Buren, former president, and Lewis Cass, Senator from Michigan. What made him attractiveanonymity plus a reputation for partisan reliability-could hardly qualify as a "strategy." • Franklin Pierce, the second dark horse, made Polk look positively ambitious, by not appearing among the candidates at the 1852 convention until the thirty-fifth ballot, when Virginia (and not his home state of New Hampshire) put his name in play. A former Congressman and former Senator, Pierce's attraction likewise lay in the lack of a policy profile in a conflictual convention with four major contenders. There was again no evidence of strategy on his part, and little room for it. • Horatio Seymour, the Democratic nominee in 1868, resembled Pierce in that he did not appear until the twenty-second ballot and was put in play not by his home state (New York) but by Ohio, while New York was actively backing Thomas Hendricks of Indiana. The attraction of Seymour came most critically from his behavior over the previous days as permanent chairman of the convention, where an even-handedness in earlier struggles among four major (and intentional) candidates made him attractive as a compromise. • While he was a first ballot nomination, the 1872 Democratic choice, Horace Greeley, former editor of the New York Tribune, was perhaps the least deliberately strategic of these four. Greeley was already the nominee of the Liberal Republican Party, one of the short-lived responses to opposition party dilemmas in the aftermath of the Civil War; he was a frequent critic of the Democratic Party in the pages of his newspaper; yet he went on to be rubber-stamped by a six-hour Democratic convention.

After Greeley, the next three presidential nominees who were not initial entrants in a nominating process were all Republicans-Rutherford B. Hayes in 1876, James A. Garfield in 1880, and Benjamin Harrison in 1888-and all were products of the internal divisions within the dominant party into three continuing factions:

• The 1876 Republican convention featured two champions for the regulars, along with the indomitable James G. Blaine for the moderates, plus a further champion for the reformers. Governor Rutherford B. Hayes of Ohio was present from the start, but only as a favorite son. He began to move on the fifth ballot when Michigan shifted to him, and he went over the top on the seventh, when both New York and Pennsylvania made further major contributions. So Hayes clearly was a strategic marker for Ohio, and must have been at least agreeable to being one. • Hayes having died, the leading contender in 1880 was former President Ulysses S. Grant, as opposed by the redoubtable Blaine and by Treasury Secretary John Sherman of Ohio, nominated by fellow Ohioan James A. Garfield. The latter did not come to life as a presence in his own right until the thirty-fifth ballot, and even then only through support from Indiana and Wisconsin rather than his home state. Yet Ohio did lead a plunge to him on the next ballot, leading to his nomination, all of which looked less like a strategy and more like a pure opportunity of the moment. • Eight years later, after the successful nomination but electoral defeat of Blaine in 1884, fourteen candidates were proposed at the Republican convention of 1888. The eventual nominee, former Senator Benjamin Harrison of Indiana, did have the support of his home state, though he was fifth among fourteen on the opening ballot. A "hidden" Harrison vote then proved to be scattered among delegates holding out for Blaine: they settled on Harrison on the seventh ballot, and he went over the top on the eighth. So while the nominee was a favorite son, he could not conceivably have gone into the convention with the strategy of hoping that Blaine would have substantial support but would not enter, and that disappointed Blaine delegates would therefore turn to him en masse.

The strategic repertoire for active nomination contenders was broadened with the arrival of the mixed system of delegate selection after 1908, since candidates now needed a conscious decision, not just about running for president and how actively to do so, but about adding one or more of what was a growing number of presidential primaries. 10 Because most of these were preference polls only, allowing no connection between expressed preferences for a nominee and the election of actual delegates, candidates generally chose not to enter. Still, there was an enhanced range of options. At the extreme, a candidate could go all-in, entering as many contests as possible. Alternatively, a candidate could enter a carefully selected subset of contests, ideally demonstrating voter attraction without risking repeat losses. Or the candidate could rest content with a home-state primary, if available, aiming to confirm an electoral attraction for the general election.

The strategy of going all-in attracted a peculiarly bifurcated clientele, made up of those who had no other hope of gaining the nomination and those who wanted just some final confirmation of what appeared to be an incipient coronation. Among those who really had little other choice, former president Theodore Roosevelt turned to this strategy in 1912, since sitting President Taft could be expected to dominate the regular party and its internal selection processes. In a different manner but with the same bleak alternative prospects, Senator Estes Kefauver chose the same strategy among the Democrats in 1952, since as an investigator of labor corruption through his Senate committee, Kefauver had no hope for autonomous support from the regular party and its main organized interest groups.

In the end, both men gained the same ultimately unproductive outcome, winning significant primaries, receiving fewer delegates from them than their vote would have suggested, and losing the nomination at the convention to candidates who had largely avoided the primary trail. On the other hand, Roosevelt in 1912 did trigger the only relegation of a major-party nominee to third place at the general election, and Kefauver in 1952 did force Adlai Stevenson to answer a second challenge from him in 1956, where Stevenson, who had consciously avoided the primary trail in 1952, went out on it in 1956 as the price of dispatching Kefauver.

Occasionally, two leading contenders, under less immediate compulsion, chose the full primary trail in a conscious effort to diminish the apparent attraction of the other. Thus Woodrow Wilson and Champ Clark entered most Democratic presidential primaries in 1912, though the result was largely a wash, and the long convention battle afterward was little influenced by these preceding contests. Dwight Eisenhower and Robert A. Taft produced much the same result from the Republican primaries in 1952: entered in most, splitting them into roughly equal parts, and seeing a short but intense convention turn on things other than preference expressions at the ballot box.

This weakness of the practical (which is to say, the delegate) side of presidential primaries was perhaps best demonstrated by William Gibbs McAdoo, former Secretary of the Treasury, in the Democratic primaries of 1924. McAdoo won almost all of the preference polls inside these contests, but he and Governor Alfred E. Smith of New York, who won no primaries at all, were to produce the most protracted convention in American party history-17 days and 104 ballots-before both lost to John W. Davis, former Solicitor General. Davis began essentially as a favorite son for West Virginia, saw his support grow through the thirty-fifth ballot, saw it decline through the ninety-fourth ballot, moved into the lead on the hundred-and-second ballot, and was nominated one ballot later.

Despite those hugely diverse roles for the presidential primarycompulsive pursuit, close division, or complete irrelevance-the most common function of the primaries was to reassure the rest of the party about the electoral potential of an apparent front-runner. Thus Calvin Coolidge, having ascended to the presidency on the death of Warren Harding, put party anxieties to rest by entering and winning the Republican preference polls in 1924. Herbert Hoover did the same with his emergent candidacy to succeed Coolidge in 1928. By contrast, their predecessor, Warren Harding, had contented himself in 1920 with winning his home-state primary in Ohio, though it was seriously contested, a result that came back to reassure supporters when he became an unforeseen compromise at an old-fashioned brokered convention.

So those were some successful cases of an essentially confirmatory strategy using the presidential primary trail. Yet even in this context and even for candidates who were not forced to undertake it and could choose their venues carefully, the strategy was hardly foolproof. Thus Governor Frank Lowden of Illinois, who did manage to stand second on the opening ballot at the 1920 Republican convention, had aspired to be the champion not just of Midwestern states but of party regulars generally. But while he did win the preference poll in his home state of Illinois, nearly uniform losses in a number of other states damaged him badly, with the result that expected support in other large organization states was divided and either arrived late or never arrived at all as those states stayed with favorite sons. 11 Despite these multiple and varied examples of life on the primary trail, the majority of all nominating contests during the entire period of the mixed system of delegate selection were not seriously contested in the primaries. Though this reticence did have a further partisan cast: Republican presidential aspirants traveled the primary route much more often than Democratic counterparts. Republicans appeared to see the value in demonstrating electoral clout. Democrats, implicitly reinforced by the additional demands from the need for a two-thirds majority, were more likely to see the down-side from primary losses. Yet at the end of the day and finally, only two contests truly turned on presidential primary outcomes, and these came right at the end of the era: the Democratic nomination of John Kennedy in 1960 and the Republican nomination of Barry Goldwater in 1964.

Kennedy, the private favorite of much of the regular party in the runup to the 1960 Democratic convention, carried major potential liabilities that were widely recognized by these same regulars. His comparative youth and inexperience were on this list; so, especially, was his Catholicism. So he needed to find some way to demonstrate that these were not serious electoral liabilities. Kennedy found the solution by defeating Hubert Humphrey, a serious, more senior, Protestant contender, first in Wisconsin and then in West Virginia, the latter believed to be hostile in principle to a Catholic candidate. In the aftermath of those wins and when none of the other contenders-Senator Lyndon Johnson of Texas, Senator Stuart Symington of Missouri, former Governor Adlai Stevenson of Illinois-were willing to take the same risks on the primary trail, Kennedy proceeded to an early nomination at the convention itself. 12 The Republican counterpart arrived four years later, when Senator Barry Goldwater of Arizona, champion of the activist wing of his party, needed a means to pry the nomination away from the great regular parties of the northeast, more moderate in ideology and more reliant on internal party structures to generate their delegates. The primaries would confirm Governor Nelson Rockefeller of New York as the electoral champion of that big, regular, northeastern wing, but in a head-to-head contest in the giant winner-take-all primary in California, Goldwater was to defeat Rockefeller, narrowly in percentage terms but massively in the associated delegate result. The convention itself would still be conflictual, with efforts to undo the emergent Goldwater majority. But California would prove to have been definitive, and a Goldwater majority going into the convention was not undone. 13 So both Kennedy and Goldwater owed their nominations in part to the availability of a primary trail. That said, the mixed system of delegate selection still allowed for many contests in the same mode as the pure convention system, where the front-runner did not triumph at the opening of the convention and the eventual nominee arose from inside the ensuing struggle. None of these were as dramatic as the instances in an older world where the eventual winner was not even nominated when the balloting began, as with James K. Polk for the Democrats in 1844 or James A. Garfield for the Republicans in 1880. Yet the point remains that if the newer world of mixed selection systems brought some true products of the primary trail, it also featured many contests on the old model, where an initial contender who was not a pre-convention leader eventually collected the nomination.

Wendell Willkie for the Republicans in 1940 was the most striking example, having held no prior public office and having entered no presidential primaries. But this account has already touched on numerous others: Warren Harding for the Republicans in 1920, entering only his home-state primary in Ohio and standing sixth among the initially nominated contenders; Governor James M. Cox for the Democrats in that same year, likewise entering only Ohio and needing forty-four ballots to achieve his (two-thirds) majority; John W. Davis for the Democrats at the following convention, entering no presidential primaries and having to endure a hundred-ballot contest between two leading opponents; and Adlai E. Stevenson as late as 1952, entering no preference primaries outside his home state of Illinois and losing that one to Estes Kefauver, yet managing to wrest the nomination away from Kefauver on the third ballot.

The coming of the plebiscitary era after 1968 would put an end to all this by altering the strategic calculations of any and all presidential aspirants. Some fundamental things did not change. The Gain-Deficit Ratio would continue to measure the impact of the bandwagon dynamic within this reformed institutional framework. The preferred occupational backgrounds and career trajectories would continue down a path inaugurated in the mixed and not the plebiscitary system, with its continuing potential for governors and a growing prospect for senators. The main structural influences would change, especially the factional structure of the individual parties, but not in coordination with the arrival of a plebiscitary system. Rather, this key shift in the factional structure of the individual parties preceded, and in that sense was part of the impetus for, a reformed era of delegate selection, while a later change in these divisions would not be accompanied by any new institutional arrangements for presidential nomination.

What would indisputably change was the near-universal implementation of state presidential primaries. Moreover, these were now to be candidate primaries, primaries that reliably linked the outcome of presidential preference polls to the creation of state delegations to the national party convention. The fall-out was to be immediate, generalized, and distinctive. Never again would the identity of the nominee be unknown when the convention assembled. Indeed, never again would eventual nominees enter the convention without already possessing their majority, though the Democratic contest of 1972, the Republican contest of 1976, and the Democratic contest of 1980 would sustain vain hopes that an obvious and ineluctable result might still be undone. After these opening years of the plebiscitary era, losers would be deprived even of this distant prospect. 14 By extension, the strategic options of presidential aspirants were sharply constrained. No candidate could any longer decide to wait for entry until the convention itself: delegates not already acquired by this candidate were guaranteed to have been acquired by someone else. No candidate could even hope to lock up his home state and then sit tight. In the new institutional and behavioral milieu, such favorite sons became anomalous, and quickly disappeared. State voters who could choose among "real" presidential contenders would prove to be thoroughly resistant to a local candidate who could not demonstrate the prospect of an impending nomination, even if they knew said candidate from home-state politics. There would no longer be anything even as conditional as a delayed nomination of the front-runner.

That said, there could still be conventions that were not unanimous on the first ballot. So losers might yet be put in nomination and enjoy their moment before the cameras. 15 But the real nomination would always have been decided beforehand, to be merely confirmed on this ballot. The original question of whether to enter the nominating contest remained, of course, but only in the sense that anyone could decide not to run. Otherwise, the new world demanded early, active, and constant pursuit of the nomination: no candidate who skipped more than the first two contests would ever again be nominated, at least as this is written. Nothing could banish death, disability, or disqualification as theoretical exceptions, but those were the only foreseeable conditions under which the winner from the primary trail would not prove to be the ultimate nominee-and they have never yet been realized.

Along the way, the question of how to enter also disappeared. The one and only option was to contest the full range of delegate selection contests, most of them candidate primaries with a few participatory caucuses thrown in. 16 The odd contest could be skipped for idiosyncratic reasons-Hillary Clinton did not campaign against Bernie Sanders in 2016 in his home state of Vermont-but those moves had to be both idiosyncratic and isolated. By subtraction, strategic considerations for the candidates were reduced to the question of when to enter, and in the modern world, even this had only three generic answers: via the Iowa caucuses, via the New Hampshire primary, or after both. Though in practice, even the last of these, while early enough in the abstract to collect a majority of delegates, would prove to be chimerical.

The roots of this modern narrowing of strategic possibilities can be traced all the way back to New Hampshire in 1916, when that state initiated its presidential primary. 17 The resulting contest was always early, becoming a local diversion in the long dark winter. But the most common result for both parties during the extended years of the mixed system was still a slate of delegates that was not publicly pledged to any national contender. This changed in the run-up to the 1952 nominations, when the state legislature amended its ballot rules to allow easier entry and, especially, to provide for an explicit candidate connection. Thereafter, New Hampshire became not just the temporal opening of the nominating contest but the first concrete indication of which candidates had votegetting prowess, a powerful lure for a mass media and a general public which reliably wanted to know what was "really happening."

That impact was magnified by the fact that this opening contest arrived, early and by itself, in a state still small enough to feature "retail" politics. What would become recognized as "wholesale" politics, relying principally on commercial advertising and supplemented by flying visits from the candidates, would become the norm for nominating politics once the delegate selection process reached the point where there were multiple primaries on the same day in larger states. But in a small state that was not sharing candidate attention on the day, retail politics, featuring inperson campaigning and direct candidate contact, remained possible. As other states woke up to the strategic requirements of a sequential plebiscitary system, the New Hampshire state legislature would take an additional step, setting the date of their contest by law at one week before the next-earliest primary, so as to maintain its initial character-as well as the attention from both the candidates and the media that went with it.

Though the surprise of the plebiscitary era, one largely accidental, was the emergence of the Iowa caucuses as the other true opening contest in this new system. 18 Iowa too had traditionally begun the process of selecting its delegates early. But with a multistage process spread over a number of subsequent months and without any candidate implications that could be readily calculated, the state had remained a backwater in candidate, media, and public attention. Yet the same reforms that were to impel an overwhelming sequence of candidate primaries also required the remaining caucus states to register presidential preferences and tie them directly to delegate selection. Iowa not only complied but actively facilitated the ability of commentators to see what had just happened at the local level and to project it through the subsequent process. And a second, regular, opening test of candidate prospects had been born.

Most candidates thereafter began in Iowa and moved on to New Hampshire. A few, expecting to do much better in the latter than in the former, skipped Iowa and waited to open their campaigns in the New Hampshire primary. A few others, with reverse expectations, opened in the Iowa caucuses but did not invest seriously in New Hampshire. Over the coming years, a variety of states would jockey to be third, and an occasional challenger would aspire to break into the top two. Yet these latter challengers reliably failed in the face of an accepted tradition plus explicit resistance by two states that were militant in defense of their privileged status, while no state that went third, fourth, or later ever came close to achieving the continuing impact of these first two. In effect, then, the strategic defaults became: Iowa-to-New-Hampshire, Iowa-but-not-New-Hampshire, or New Hampshire-but-not-Iowa. Table 6 .3 compares the pay-off from these strategies, results that can in turn be made to tell a set of lesser stories, mixing the possible strategies and outcomes: • At one extreme, winning in Iowa and New Hampshire (1st in IO + NH) was tantamount to nomination. However long they managed to stagger onward, all other candidates were dead after New Hampshire under these conditions, and the contest was effectively over. Two opening wins was, however, an uncommon result, perhaps because Iowa and New Hampshire were so very different demographically, perhaps just because participatory caucuses tend to sample their electorates so very differently from primary elections. • At the other extreme, winning in Iowa (1st in IO) or New Hampshire (1st in NH) were not by themselves especially portentous, given that these were reliably the two contests with the largest number of contestants, that is, the largest number that would prove to be irrelevant. A candidate who won one or the other was more likely to be nominated than a candidate who did not win either, yet there were so many more of the latter than the former, given the usual raft of aspirants who would prove to be going nowhere, that a single win was not all that strong a prediction of ultimate nomination. • Though combing the two together, that is, winning one or the other by comparison to losing both, was again portentous. The collection of Democrats and Republicans who won at least one of the two (1st in IO or NH) was overwhelmingly likely to contain the ultimate nominee. An early win more or less automatically improved name recognition while simultaneously suggesting electability. At the same time, it created a veritable field of losers, many of whom had invested all their resources in these opening contests, to the point where most would never go on to win anything. Anyone who then demonstrated an inability to win another contest quickly-even though there were more than forty yet to come-would have a very difficult time attracting fresh activists, fresh donors, and fresh voters by comparison to the candidates who were in fact winning. • Before Joe Biden captured the 2020 Democratic nomination, it could also have been said that no candidate who had not come in first or second in Iowa or New Hampshire (1st or 2nd in IO or NH) had ever been nominated in the plebiscitary era. In that limited sense, the Democratic contest of 2020 was precedent-breaking. The critical contest, the South Carolina primary, actually did come after Iowa and New Hampshire, while the winner, Joe Biden, had not placed either first or second in these prior contests. Yet even Biden could not break the apparent behavioral rule that every ultimate winner had to have contested one or the other of those first two delegate selection events.

Rudolph Giuliani, former Mayor of New York City, did produce the essential ultimate contrary test. Leading in the polls in 2011 but lagging in both Iowa and New Hampshire, Giuliani gambled on waiting until the Florida primary, fifth in the sequence in 2012, where he was leading when his campaign decided to pass on the first two contests. Yet the interim was to see John McCain, Senator from Arizona, win the New Hampshire primary and Mitt Romney, former Governor of Massachusetts, place second in both New Hampshire and Iowa. So twenty-six days after Iowa and twenty-one days after New Hampshire, Giuliani found himself opening his campaign in the face of the apparent front-runner in McCain and the anointed challenger in Romney, while he himself was to finish an undistinguished third in Florida, garnering a vote less than half that of either of those others.

Those are the key strategic decisions facing-and inevitably made bythe individual aspirants for a presidential nomination across time. Some of these aspirants had little choice in the matter: individual decisions were more or less made for them by their own inherent constraints. Others had more leeway, being better-connected or better-resourced, though by the coming of the plebiscitary era, even the best-off had little ability to reshape the overall course of nominating politics on their own. Yet there was another set of strategic decisions that interacted with these candidate calculations from the other side, a set that belonged not to the presidential aspirants, individually or as a cohort, but rather to the individual states and their state parties. As decisions, theirs were more episodic than those available to the candidates, but simultaneously more fundamental in that they tended to be larger in their reach and more lasting in their impact.

These state choices put the details into the specific course of a nominating conflict, and even when the national parties were willing to impose some nationwide standards and restrictions, much remained largely the province of the states. Their decisions involved state-based choices for the institutions of delegate selection, state-based systems for delegate allocation, and state-based efforts to maximize their own influence within the unfolding collective sequence of delegate selection contests. In turn, state decisions inevitably interacted with the strategic choices made by individual candidates. It is that sense in which the state side of nominating politics fills in the specifics of any given contest, to the point where the two influences together-individual campaign strategies coupled with state party adjustments-sometimes went so far as to provide the final shove favoring one aspirant over another.

These alternative decisions for the states and their parties once again differed by institutional era. In fact, one major instance-one major strategic choice and hence one key process-characterizes each temporal period. Under the pure convention system, in the years from 1832 to 1908, the main option available to the individual states involved applying the unit rule to their respective delegations or not. During the mixed system for delegate selection, in the years from 1912 through 1968, the main option was instead whether to adopt a presidential primary or stay with a traditional convention, and how to structure a primary if they made that choice. And in the modern plebiscitary era, 1972-2020 and counting, the major remaining choice was where to situate their state in the long chain of presidential primaries and occasional party caucuses through which nominating politics ran-and where the combination of these state choices was what produced the actual sequence of delegate contests.

The pure convention system featured a general uniformity in the intraparty mechanics of delegate selection across the states. Still, there were some further but lesser options in the detailed workings of selection politics among the individual parties in this long opening period. Thus states differed in the locus of ultimate selection. Some used regional party gatherings plus a further increment from the state convention; others assigned creation of the full delegation to the state party in convention; and some went on to reserve the option of selecting the full delegation to the state central committee, without so much as a formal convention. Yet the main option for state parties in this opening era was whether to apply a unit rule to any given result.

States with small delegations tended not to do so. From one side, interpersonal relations could handle the proceedings of the delegation. From the other side, those relations could be actively damaged by compelling the behavior of colleagues who were personal acquaintances and whose coordinated contributions might well be needed for other party business. So the states that did use this further power of constriction tended to fall into two categories:

• First were the big states with classic machine-type parties, found mainly in the northeast. More than the others, these states arrived at the national party convention with the conscious-and apparently plausible-intention of maximizing their influence over the ultimate result. The unit rule was one more potential tool in pursuit of that goal. • The other reliable users were the southern states. Their results in November were known in advance, for which southern Democrats wished to be rewarded and southern Republicans excused. Yet they also had policy concerns specific to the region, and both Democratic and Republican parties could hope to use the unit rule to pursue these.

But how much did it matter? That is, how good were either of these state clusters at securing the outcomes they desired-or desired to avoid? And more consequentially for nominating politics, how much did this record of success or failure alter an ongoing bandwagon dynamic, the shifting backgrounds of nominees, or the influence of party factions?

In the case of the major northeastern states, the crucial test is perforce focused on the three outsized examples, New York, Pennsylvania, and Ohio. Though among Republicans, a tour of national party conventions in the pure convention era suggests that practical opportunities to apply the unit rule were more limited than its theoretical availability would suggest.

Before 1876, the Republican Party would in principle tolerate imposition of the unit rule convention-wide; after 1876, implementation was purely a state matter. Yet there was no uniform imposition for the three major states-and no uniform success when the rule was imposed. New York did apply the rule in 1860 at the first serious Republican convention, yet in doing so it stayed with Senator William H. Seward, its favorite son, even through the final ballot when the vast bulk of the others shifted to Abraham Lincoln. New York had a favorite son and stayed with him again in 1876, until Rutherford B. Hayes became the only remaining option to James G. Blaine. After that, the state was normally so divided as to make imposition impossible. New York did move early as a bloc to Benjamin Harrison, the ultimate nominee, in 1888. Yet it was also the lone major hold-out against Harrison when he was renominated in 1892.

At the opposite extreme, Ohio was equally devoted to its favorite sons, but more successful in using procedural weapons to support them. The Buckeye state possessed a favorite son in 1860 but nevertheless broke to Lincoln on the third ballot, a break that was important to his nomination. Ohio stayed with its favorite son in 1876, Governor Rutherford B. Hayes, which gave him support sufficient to survive until the convention was ready to seek a compromise nominee. Ohio likewise stayed with Treasury Secretary John Sherman, yet another favorite son, against the Grant renomination effort in 1880, stalling the Grant drive until a second Ohioan, Congressman James A. Garfield, came into play. It lost this uniform winning record in 1888, when it stayed with another favorite son until the nomination of Benjamin Harrison of Indiana was an inescapable result. And it came back to stay with home-state Governor William McKinley in 1892, even as Harrison was overwhelmingly renominated.

Lastly, Pennsylvania had an even longer but less imposing roster of favorite sons across this period, yet it ended up with the winner, James G. Blaine-not a favorite son-only in 1876. Otherwise, imposition of the unit rule on behalf of also-rans from Pennsylvania resulted in a reliable trip to the political wilderness. For the three states as a group, then, there was only intermittent application of the rule. Yet when it could be imposed, the record of strategic pay-offs was spotty at best. New York had the hardest time of the three in imposing the rule at all. Pennsylvania had the easiest time but the worst pay-off. So only Ohio wielded this strategic power with real, albeit still intermittent, impact on the eventual nomination.

In theory, the skeletal Republican parties of the south during this period might have resurrected the rule on a regional basis. But in practice, the southern states were often split in their judgments of which candidate would best provide the essential federal patronage that kept these skeletal parties alive, while individual delegates were further split over their prospective personal advantages. So imposition of the unit rule was comparatively uncommon. The exceptions were 1880, where the possible return of Grant held out the hope of bringing back a patronage heyday for regional Republicans, and 1884, when these states provided the essential base, albeit another unsuccessful one, for the effort by President Chester Arthur to secure renomination in his own right.

On the other hand, the unit rule was always more of a favorite among Democratic parties, where its clout was potentially magnified by the need for a two-thirds nominating majority. In the antebellum years, the bigthree states were unanimous for Van Buren in his attempt at a third nomination in 1844, before all three broke to Polk, the inaugural dark horse. The three were again united in imposing the rule for 1848, but two of them, New York and Pennsylvania, went with favorite sons and not with Lewis Cass, the nominee. Ohio and New York were both split in 1852 and 1856, while Pennsylvania applied the unit rule on behalf of its favorite son, James Buchanan, unsuccessfully in 1852 but successfully in 1856. Conversely, Ohio and New York applied the rule in 1860 on behalf of Stephen Douglas, the ultimate nominee, while Pennsylvania was the one that was split throughout.

So the big northern states were more often able to apply the unit rule in this period among Democrats as opposed to Republicans, though their record of success in return was equally spotty. After the Civil War, the main interparty decisions for the Democrats involved the attempt to find a nominee who might restore the national party to competitive status, which quickly focused convention politicking on the fortunes of any available reform governor from New York. Yet as consensual-and successful-as that particular strategic judgment turned out to be, the bigthree states, with presumably a great deal to gain, repeatedly splintered in response.

A unanimous Ohio delegation in 1872 contributed the major shift on the twenty-second ballot that started the bandwagon for Horatio Seymour, former reform governor of New York. In 1876, it was instead New York that delivered a unanimous vote for Samuel J. Tilden, sitting reform governor, while both Ohio and Pennsylvania sat with their own favorite sons on the critical opening ballot. In 1880, the three went their separate ways, with New York floating a pair of possible successors to Tilden, Ohio remaining behind its favorite sons, and Pennsylvania split and marginal to the nomination, despite the fact that Winfield Scott Hancock, the eventual nominee, was a Pennsylvanian. New York again turned to the unit rule in 1884 to support Grover Cleveland, while Ohio was split and Pennsylvania opened with a favorite son, before both became part of a general move to Cleveland on the second ballot.

The Democratic convention of 1896 registered the first coming of the populist revolution to the national Democratic Party. This time, a strong dissenting view from regular parties in the northeast actually unified the big-three states. Yet imposition of the unit rule on top of a common cause would still leave all three on the losing side for a generation. All three big-state parties were again in vigorous opposition as the populist impulse steamrolled the Democratic convention of 1896, with both New York and Pennsylvania bootlessly invoking the unit rule first in support of the gold standard and then against William Jennings Bryan as the nominee. Bryan was so evidently unanimous as the (re)nominating choice in 1900 that none of the big-three bothered with an even more bootless opposition. Though when the opportunity returned to vote against Bryan at the 1904 Democratic convention on procedural issues and against populism via alternative champions, New York, Pennsylvania, and Ohio all stayed with the unit rule and with New York Justice Alton Parker, the lone traditional conservative during the 1896-1908 stretch of Bryan nominations.

By contrast, most southern states had an easy time imposing the unit rule in these antebellum years, where an underlying homogeneity as a regional bloc could be focused on the one particular candidate who was ordinarily judged most sympathetic to the region as a whole. Lewis Cass in 1844 and 1848, a classic northerner with southern sympathies, was easy. The southern states remained en bloc in support of Franklin Pierce, rewarded for his regional neutrality as president in his unsuccessful renomination bid in 1856. And by 1860, the real question was neither the unit rule nor regional unity, but simple southern continuation in the national Democratic Party, as Alabama, Arkansas, Florida, Louisiana, Mississippi, South Carolina, Texas, and Virginia each walked out at one or more points during that extended and disastrous convention. 19 After the war, the southern states, more isolated than ever as a region, were often content with a reform governor from New York as the way to appease three main party factions: southern loyalists, northeastern immigrants, and western populists. Yet they usually added a wrinkle to their intraparty strategy by rallying to an explicitly southern champion first, before moving to an ultimate compromise. Thomas F. Bayard, Senator from Delaware, was the most common vehicle for this, and he attracted a small, largely southern vote at the 1876 convention. He reappeared as the more conscious instrument of a "southern strategy" in 1880, when Bayard placed second to Grover Cleveland on the opening ballot. And he appeared once more, standing second on the opening ballot in 1888, before the southern states came back into line behind Cleveland. 20 That particular strategy expired with the convention of 1896, when the rise of populism and the accompanying candidacies of William Jennings Bryan proved comfortable to these southern Democratic parties. The unit rule continued to be generally imposed across the region, but a generalized regional preference for candidate Bryan meant that this had little separable impact on the bandwagon dynamic, candidate backgrounds, or major structural influences. In sum, the south as a region was even more likely to impose the unit rule than were its big northeastern brethren, but it was harder to separate the unifying effect of this imposition from the unifying impact of the search for a candidate who would protect or advance common southern interests.

The arrival of the presidential primary in 1912 was to be the opening move in what would become two, not just one, extended institutional regimes for delegate selection and presidential nomination:

• The first of these became the mixed system of delegate selection, stretching through the half-century from 1912 to 1968. It was "mixed" because state parties received their first major institutional option since creation of the national party convention itself in the 1830s. In turn, some states did indeed avail themselves of the choice, though most were to stay with the traditional state party convention.

• A second institutional regime displaced the mixed system after 1968, morphing quickly into the modern plebiscitary system and itself extending through the next half-century. In principle, states could choose between a reformed presidential primary or a reformed state convention. But in practice, the primary would become so dominant as to justify calling the result a plebiscitary era.

In the beginning, however, the states gained a major strategic option, most aspects of which would actually be taken away in the successor period. Yet during what became the long mixed period for delegate selection, the appearance of this new institutional option was to have two general but contrary effects on state strategies. From one side, state parties could now choose a whole different mechanism for delegate selection, while simultaneously tickling its details to suit local preferences. 21 Yet from the other side, the practical impact of new institutional arrangements would remain limited and intermittent. Most states in fact experienced no change at all: the bulk of national convention delegates would continue to be chosen by way of established-internal-party machinery through 1968.

Still, the mixed system of presidential nominations had arrived, courtesy of those states that did move to a presidential primary, and the minority that did adopt the new primary was large enough to make the composite institutional matrix look different. Within its boundaries, states could choose not just between the new primary or the traditional convention to create their national convention delegates. Those who switched to a primary could choose additionally among three variants: one linking a presidential preference poll directly to delegate election, one offering both a preference poll and delegate election but without an automatic connection, and one focusing on direct election of the delegates, absent a presidential preference poll.

A few states injected not just the letter but the spirit of nominating reform into their institutional processes. This implied not just a presidential preference poll on the public ballot, but a link between this poll and the selection of individual delegates. Wisconsin was an early example, adopting this form of (true) presidential primary at its first opportunity in 1912 and retaining it through 1968. Though Wisconsin was simultaneously good at reminding analysts-and presumably the candidates themselves-that primary voters did not have to opt for an obvious national contender, and that when they did so, there could still be substantial flexibility in the link between preference poll and delegate result.

In fact, in a first test in 1912, Wisconsin Republicans went on to cast 73% of their newly empowered vote for a favorite son, Senator Robert M. LaFollette, Jr., apparently giving him all the delegates, though these delegates would ultimately abstain from voting on a presidential nominee at the convention itself. That same year, Wisconsin Democrats were to favor Woodrow Wilson over Champ Clark, giving him 56% of their vote but 77% of their delegates. 22 Fifty years later, Wisconsin Republicans would give 80% of their preference votes and 100% of their delegates to Richard Nixon, while Wisconsin Democrats gave 56% of their preference votes along with 83% of their delegates to Eugene McCarthy.

More states offered both a presidential preference poll and direct election of their delegates, but on a ballot that provided no automatic connection between the two. In principle, an aggressive nominating campaign on the part of a presidential contender could forge that connection. But for an out-state contender (which most were) who was not widely known in-state (as most were not), this was a serious operational challenge, made only more difficult by the fact that such campaigns were implicitly challenging the existing party organization. Pennsylvania was an early proponent of this system, adopting it in 1912 and continuing with it through 1968.

In its initial round for the Republicans in 1912, Theodore Roosevelt gathered 60% of the preference poll in the Keystone state and 84% of state delegates, while Woodrow Wilson was the sole entrant in the Democratic preference poll and came away with 93% of the delegates. In the final round of this arrangement in 1968, Richard Nixon won the Republican primary on a write-in, scoring 60% of the vote, yet Nelson Rockefeller acquired 41 of the 64 state delegates, for 64% of the total. At the same time, Eugene McCarthy won 56% of the official Democratic preference poll, while Hubert Humphrey acquired 104 of 130 delegates, 80% of that total, despite having no one on the ballot who was publicly committed to him.

Lastly, some states, including New York with the largest delegation during this entire period, dispensed with the preference poll and simply offered the names of all contenders for delegate status. Again in principle, individual candidates for national convention delegate were free to publicize their preferences, though that was easier said than done by any one individual. More commonly, voters chose delegates based on general public recognition or ongoing party connection. New York stayed with this arrangement from the first moment of the mixed system in 1912 through its final moment, or in some sense past the final moment, since New York did not initially comply with Democratic reform rules passed for the 1972 contest.

Absent a preference poll, it was only possible to say that William Howard Taft for the Republicans got 76 of the 90 New York delegates in 1912, while Champ Clark, thanks to imposition of the unit rule, received all 90 of its Democratic delegates. A half-century later, Nelson Rockefeller would acquire 88 of 90 delegates for the Republicans, while Hubert Humphrey would outgain Eugene McCarthy for the Democrats, 97 to 87, despite not announcing his campaign for president until the calendar deadline for putting delegate names on the ballot had passed.

That institutional world would be worth more than a half-century, but would ultimately be swept away by the wide-ranging reforms of the 1970s, almost immediately in the case of the Democrats, more gradually in the case of the Republicans, yet swept away in both cases. The result was the plebiscitary era of presidential selection, beginning in 1972 and heading for another half-century as this is written. The new era would be characterized most simply by the overwhelming dominance of a sequence of presidential primary elections. Yet this overall dominance risked obscuring the real engines for the constriction of institutional choice inside the overall shift, engines that involved the type of presidential primary that a given state could utilize or the type of party caucus it could retain if it chose not to go with the drift toward primaries.

The overall move away from internal convention-based systems and toward external primary-based systems was dominant, self-evident, and overwhelming 23 (Table 6 .4). For a very long time beforehand-really the entirety of the mixed system of delegate selection-there had been only individualistic further fluctuations in the adoption of the presidential primary. Primaries had burst on the scene in a major way in 1912, giving the succeeding era of nominating politics its "mixed" character. Yet from there through the end of the era in 1968, there was only random further variation: a primary added here, a primary subtracted there, but a long-run picture of impressive stability. Measured by the simple presence or absence of presidential primaries, 1968 looked more or less indistinguishable from 1912. 24 Table 6.4 Number of  presidential primaries  across time   Dems  Reps   1912  12  13  1936  14  12  1952  16  13  1968  15  15  1972  21  20  1980  34  34  2000  40  43  2016  39  39  2020  41  41 That was instantly changed by the first big round of Democratic reforms in the run-up to 1972, with an institutional fall-out reaching into both parties. So by comparison to the long run of institutional stasis over the previous fifty-six years, the scale of change was already noteworthy. Yet within two further iterations, that is, by 1980, about two-thirds of the state parties on both sides of the partisan aisle had gone to the primary. Moreover, this aggregate of states understated the dominance of the shift, since larger states were more likely to make the move. By 2020, about forty of the fifty states could be expected to use primaries rather than caucuses, selecting close to ninety percent of the delegates.

Nevertheless, that still hid the larger half of the institutional impact that came with-but really created-a new plebiscitary era. For in fact, the nature of the presidential primaries inside this burgeoning primary gain, coupled with the nature of the party caucuses that remained among states that did not move to the primary, brought the larger share of the practical impact from overall change, rendering the composite result truly and strikingly different 25 (Table 6 .5). In the previous world of the mixed system, delegate primaries, those that elected their national convention delegates without an explicit presidential attachment, far outnumbered candidate primaries, the ones where presidential attachment produced the delegate result. Just as traditional conventions, with their essentially internal selection mechanics, solidly outnumbered participatory caucuses, where open public meetings began the process instead.

The balance among these four institutional arrangements-along with an occasional selection limited to a state central committee-looked essentially as it had since 1936, the year the Democrats finally eliminated the two thirds rule, up through 1968, the effective end of the mixed system of delegate selection. Reconstitution was clearly underway by 1972 when, if the generic primary had not yet fully dominated the composite process, the new additions were all candidate primaries, and in both parties. By 1976, after only one more iteration, the delegate primary was nearly extinct, salvaged among Democrats only by the state of New York. Both the delegate primary and traditional party conventionalong with those stray committee selections-would die off entirely within the decade, by which time the dominance of the candidate primary had nearly quadrupled among Democrats and more than quadrupled among Republicans. State parties that wished to maximize their individual clout in nominating politics were effectively left with two remaining options. The first involved the choice between a candidate primary and a participatory caucus, where the choice of the latter could be important to the individual state but was truly minor-with the great exception of Iowa-inside the larger process. The second involved the temporal placement of any individual contest, which requires a separate section below. Here, note merely that, as the lone recurrent option remaining to state parties in the modern era, it would be more frequently attempted.

The first option, the choice between a candidate primary and a participatory caucus, was institutionally fundamental. The two institutions not only created their delegates in entirely different fashions; they were even linked to the general public through differing routes. That said, there was never much actual competition between the two. The tide of reform in the 1970s and after, propelled by a desire to increase public participation in party affairs, ran heavily in favor of candidate primaries rather than participatory caucuses. 26 Primaries reliably involved far more individual citizens than caucuses, while potential voting publics were far less attracted by the argument that the selection process as a whole would be improved if they themselves were removed, or at least sharply reduced, by means of a participatory caucus.

The institutional outcome was clear enough. Not only did the number of states turning to the primary continue on an upward path. The comparative size of the states utilizing primaries continued to rise as well (Table 6 .6). Already by 1976, the states that had turned to candidate primaries for delegate selection contained more than two-thirds of all Electoral Votes; only Texas among the larger states had not joined the parade. By 2016, forty years later and deep in the plebiscitary era, this Electoral Vote imbalance, two to one at the start, had reached eight to one. 27 No major states and very few of the middle-sized remained with the participatory caucus. So the caucus effectively relegated its remaining users to spectator status within nominating politics, again with the dramatic exception of Iowa.

The lone compensation to the lesser rump of small states that still clung to the caucus option, beyond managing to keep delegate selection focused on party activists and out of the hands of their rank and file, was that the delegate reward per caucus participant was far greater than the delegate reward per primary voter 28 (Table 6 .7). Which is to say: it required far 

The major option remaining for state parties that were trying to maximize their nominating clout involved the choice of an individual date for what was now almost always a candidate primary. In principle, this choice provided substantial degrees of freedom, in that individual states could choose to go early, middle, or late within the sequence of state primaries as a whole, while calibrating this placement additionally by considering the clustering of the other states, though their temporal strategies were simultaneously in flux. Yet the cumulative impact of such moves, widely attempted and constantly adjusted though they were, would prove much smaller than the often-excited predictions of reformers, strategists, commentators, and participants. To begin with, the more a reformed process encouraged states to move their delegate selection contests, the more the simultaneous collective maneuvering that resulted tended to cancel out these individual efforts.

More of this underwhelming return on placement strategies was due to the fact that one of the three theoretical options-going earlywas largely foreclosed: the successful assertion of temporal primacy by Iowa among the caucuses and New Hampshire among the primaries had severely reduced the possible impact of going early on the part of everyone else. Nevada did increasingly succeed in asserting a right to be the second caucus, and South Carolina did increasingly succeed in asserting a parallel right to be the second primary. 29 Yet the shaping influence of Iowa and Hampshire was so overwhelming in terms of campaign investment and media attention, and so reliably determinative in narrowing the field of presidential aspirants, that these next two states, even together, had little hope of a consistent comparable impact. 30 Twenty-two aspirants qualified to participate in the televised Democratic debates before the first actual contest in 2020; exactly five were even vaguely plausible as nominees after Iowa and New Hampshire.

Indeed, the impact of the two opening contests was ordinarily sufficient to shift the focus of nominating campaigns away from wins or losses in individual states, especially small ones, and toward the raw accumulation of delegate totals, which made large states or large clusters the central target for presidential aspirants. In the process, room for strategic maneuvers by way of the choice of a temporal location had effectively been compressed to "middle" versus "late," though conversely, what was the middle of an evolving post-reform sequence would creep forward over time. Which is to say: a growing consensus on choosing middle rather than late would produce a generalized push toward "front-loading," a pressure that only increased over time while simultaneously reducing any incentive to wait. Though at the start of the plebiscitary era and for some elections to come, the logic of going late retained some real attraction.

The contests of 1972, the first ones after sweeping party reform, still featured a hybrid quality in this regard. From one side, most states that had previously possessed primaries kept them in the same temporal location but merely switched, as necessary, from delegate to candidate format. From the other side, states that were introducing a new primary tended to look at what appeared to be the available-that is, the open-slots within the existing calendar. 31 The initial result for 1972 and the next two contests kept the largest single aggregate of delegates at the end of the selection process (Table 6 .8 A and B). Though in truth, this was not so much an aggregation of multiple states as a register of strategic decisions by a few states with large delegations. For them, the logic still seemed simple: late contests could be the ones that chose among the serious remaining contenders and crowned a specific nominee.

The 1984 nominating contests would break this mold, setting off what became a chain reaction of strategic choices, leading to a clearly different overall sequence of state contests and eventually to the much-debated front-loading (Table 6 .8 A and B). For 1984, the immediate impetus was an attempt by southern states within the Democratic Party to coordinate their primaries in order to maximize regional influence over the identity of the Democratic nominee. 32 On its own terms, this effort was not successful: only three states conformed to what was mooted as an ostensible region-wide strategy. Though in passing, their moves drew the attention of several non-southern states, so that the upshot, for the first time ever, was the single largest concentration of states and delegates in an earlier temporal location, March 13, 1984 . The pattern would be sustained through 1988 and 1992, with the largest concentration of delegates available in early March and the secondlargest concentration available in early June. Though by 1988, the early concentration had become massively the largest concentration of primary delegates ever seen on one day in the selection process: 38% among the Democrats and 32% among the Republicans. The two nominating contests of 2000 then altered the second piece of these patterns, by moving the main secondary concentration of delegates away from the end of the process and far forward. In that year, Super Tuesday was on March 7 for both parties, but the second main clustering was now March 14, exactly one week later, taking with it any serious support for the strategy of going late.

This secondary cluster then remained in March through 2008, but the main cluster continued to shift forward, getting to February 5 for both parties in 2008. Both national parties would eventually respond to this "calendar creep" by specifying limits within which states could and could not locate their primaries, on pain of having the credentials of their delegates rejected at the national party convention. 33 In response, the Republicans in 2012 and the Democrats in 2016-the Democratic nomination in 2012 was uncontested-were to see Super Tuesday migrate back to early March, a bit later in contemporary terms but still very early by historical standards. That date continued to capture the largest single cluster of state parties, along with the largest single cluster of convention delegates, though these were below totals for previous years. In the process, the old secondary cluster of available delegates-toward the end but not at the very end this time-would be restored as well.

Along the way, the evolution of this front-loading proved sufficient to produce a feedback effect in Iowa and New Hampshire, driving them to move the two benchmark "early" contests earlier and earlier (Table 6.9). New Hampshire was the lead actor here, shifting its primary in 1976 from the traditional mid-March location to late February, in sync with the arrival of the plebiscitary era; then to early February or late January from 2000, as front-loading took hold across the party; then to early January in 2008-before the national parties took control of the entire calendar. A new, national, and explicit counterattack on calendar creep Iowa had made its original-and even more consequential-contribution to this process in 1976, the year it switched to revealing candidate preferences in what as a result became the opening contest in the entire nominating sequence. At the start, this put Iowa a full month ahead of New Hampshire, but from 1984 onward, under informal pressure from the national party to move its contest closer to the total sequence, it settled on being essentially a week-eight days-before the date of the New Hampshire primary. Thereafter, aping the positioning of New Hampshire, Iowa would move to early February, then to late January, then to early January-all the way forward to January 3. At that point, the same national reconstitution of 2012 that moved New Hampshire back a month would move Iowa back a month too, to early February.

But once again, how much did all such maneuvering matter? The coming of the plebiscitary era did reconstruct the institutional context for delegate selection, eliminating the traditional party convention along with the delegate primary, while replacing them with a sequence dominated overwhelmingly by candidate primaries. And inside that sequence, the strategic gambits of individual state parties went on to rearrange the ordering of these selection contests and, even more crucially, its clustering. Yet Chapter One confirmed that a bandwagon dynamic with its diagnostic measure, the Gain-Deficit Ratio, had characterized nominating politics since the 1830s, burning across three sharply different institutional frameworks. So the coming of the plebiscitary era did not alter, much less derail, an ongoing dynamic, while reorganizing the sequence and clustering of contests inside newly plebiscitary institutions was even less likely to achieve any such effect. Table 6 .10 demonstrates this basic continuity despite all such state party maneuvering, by offering the specific points when the Gain-Deficit Ratio was achieved in all twenty-eight of the nominating contests in the modern era. These are the points when the nomination was effectively settled, aggregate vote totals notwithtanding and regardless of contrary interpretations by candidates and commentators. Seen this way, the selection of an aspiringly strategic date for one or another state primary looked The collective impact of individual state efforts at the strategic placement of their contests did alter the formalistic sequence of delegate selections, most obviously by moving the date of the largest cluster of available delegates, ranging from June 8 at the latest to February 5 at the earliest. Yet that did not prevent the bandwagon from running in a parallel and familiar fashion across all twenty-eight shifting contests in this modern era. By comparison, individual state calculations, or even of the collective front-loading that followed from it, were inconsequential. Said the other way around, the impact of front-loading on the bandwagon dynamic was exiguous.

As ever, this did not mean that the details by which each contest unfolded were the same, nor that the institutional reordering of these contests had no effects. The bandwagon did role later when the bulk of the delegates were selected later, and earlier when they were selected earlier. Though even within this mechanical impact, there were years when later clustering meant later effective nomination, as with 1972 and 1976. Just as there were years when earlier clustering meant earlier nomination, as with 2000, 2004, 2008, and 2020. Yet there were also years, the modal category, when the ratio was achieved after the date of the largest delegate cluster, as with 1984, 1988, 1992, 1996, 2012, and 2016 . And there was even a year when the ratio was achieved before that date, in 1980.

That composite picture underlines the inefficiency of individual state strategies as a collective whole. And the associated irony is evident. If states remained where they had been historically, they quickly discovered that their contests had become irrelevant. But if states sought a fresh advantage from temporal placement, this meant that the process would become increasingly front-loaded, with the result that individualized gains for particular states within this augmented clustering could only get smaller and smaller, not larger and larger. But were there no individual successes within this largely self-canceling collective failure?

The prospects of any such effect among the smaller states brought another immediate irony. The statistically inconsequential states of Iowa and New Hampshire were the two states of any size that captured in practical politics what they lacked in raw numbers. But once they had secured the only temporal positions that could make up for their small size, by creating, defending, and harvesting their opening locations, Iowa and New Hampshire doomed other small states to disappear among their larger counterparts. So if there was to be some strategic effect elsewheresome individualized triumphs buried within a collective failure-it could come in only two places. In the first, it had to come from coordination among a bloc of mid-sized states with some common (unifying) characteristic. Or it had to come from the largest states once again, the ones with some numerical hope of manipulating the total process on their own.

The best example of the bloc strategy-albeit ultimately just a great new instance of self-defeating efforts-surfaced in 1984 and reached its apogee in 1988, when a growing group of southern states attempted to create a regional cluster that would wield more practical influence than any of them could generate individually. After a halting effort in 1984, the southern states largely succeeded in moving as a group to what became Super Tuesday in 1988, injecting twelve southern and border states into that one day. Yet when these newly clustered states actually voted, their strategic gambit fell afoul of two inherent defects. In the first, even the south proved to be much more heterogeneous in its candidate preferences than strategists had imagined, allowing a range of candidates to survive comfortably: Senator Al Gore of Tennessee, the moderate leader in the field; the Rev. Jesse Jackson of Illinois, the most liberal; and Governor Michael Dukakis of Massachusetts, the ultimate nominee. And once the bulk of southern states had voted, the south as a region was of course devalued in the remainder of the contest, having few regional representatives left to register their preferences (Table 6 .11). The big-state story was different, but ultimately came with the same self-defeating moral. Commentators at the time did not fail to notice the onward march of front-loading; indeed, they anguished over it. What they did fail to notice was that this new phenomenon, this front-loading, was actually caused by the behavior of the cluster of large states that had kept the largest delegate cluster anchored at the end of the nominating sequence for a long time. When these states finally abandoned that position, recognizing the growing importance of a couple of early delegate clusters and moving to join them, it was in fact this move that converted a statistically prominent date, Super Tuesday, into a frequently decisive one, while simultaneously crushing the individual efforts not just of the smallest states but also of aspiring blocs among the middle-sized.

California, New York, and Ohio could take the most immediate credit for this effect in 2000 and 2004, while California, New York, Illinois, and New Jersey took the result to its zenith in 2008, the year when 48% of the delegates in both parties were available on Super Tuesday. But the key fact inside all this was that these three or four states were more than half of the delegate concentration on that critical date in all three years. There were still 11, 7, and 8 other states located on this date in 2000, 2004, and 2008 respectively, so it would have garnered the largest number of states without the contribution of a handful of giants. Yet those remaining states had become small change once these giants moved in. So the lead story-and the essence of the new dynamic-was the behavior of the big states that had previously anchored the final stage of the selection process.

For a student of the impact of state strategies, the upshots were cruel. Once again, the collective impact of a set of state maneuvers proved to be largely self-canceling. The mid-sized states that had begun a coordinated move toward the front of the plebiscitary sequence, in order to escape the domination of Iowa and New Hampshire on the front end and of a cluster of giants at the back, did set off the creation of a couple of greater and greater (and earlier and earlier) delegate clusters, the largest of which was usually dubbed Super Tuesday. Initially, this was an improvement, albeit a limited one because these states were not nearly as homogenous as strategists had imagined. Yet any such advantage dissipated quickly when the largest states, recognizing the new delegate distribution inherent in this new sequence of state contests, move to join the couple of major dates with rising influence over the nomination-where they simultaneously became the explanation for this rising influence and the crucial elements within these new delegate clusters.

What was not too early to perceive was the continuing inability of even the largest states to use their strategic possibilities to overcome, alter, or so much as divert the behavioral logic of a long-running bandwagon dynamic. In that sense, analyses of the impact of the rules of the game and analyses of the strategies adopted in response to them, by state parties and presidential candidates, were very similar. Rules reform, albeit very intermittent until the modern era, did have real and measurable impacts on the apportionment of delegates among the states and on the allocation of those delegates among presidential aspirants. It was just that these impacts were reliably minor, when they were not actually contrary to the intentions of rules reformers.

Adding strategic impacts to this picture changes its specifics but not its overall conclusion. From the earliest days, presidential aspirants did have to decide whether to run for a presidential nomination, when to enter the nominating contest if they wanted to run, and how to proceed once they had answered the first two questions. Yet their realistic options were powerfully constrained by their own assets and liabilities and then by their indifferent success at gaming an actual nominating contest, as diverse as these contests reliably were and where most of the other contestants were weighing the same strategic considerations. As a result, most strategies were unsuccessful in the ultimate sense of securing a nomination, while many that looked like successes had to be considered largely accidental.

One of the biggest things that serious strategists had to consider was the specific course of a nominating contest, that is, the temporal placement of state contests within the overall process, the way one contest might (or might not) leverage another, and the potential collective impact of the resulting sequence. This aspect of their strategic environment did become more varied, or at least more complex in its detail, as time passed. Moreover, state parties were not just aware of being integral to the strategies of potential nominees. They worked constantly to adjust their own place within the nominating sequence so as to maximize state influence over it. With the result that the individual pieces were often in motion, while the sequence itself changed regularly, courtesy of those changing attempts at state influence.

Yet candidate strategies proved to be more constrained in practice than an abstract description might suggest, in part because the sequence of nominating contests brought its own constraints, in larger part because the bandwagon dynamic easily absorbed a changing sequence. Beyond that, state strategies proved not only to have inherent constraints of their own-leeway for strategic placement was effectively quite narrow-but often to become self-canceling when any significant minority of states, or just a small set of states with a significant number of delegates, reached similar conclusions about how to game the overall contest. So in the end, strategies and sequences told essentially the same story as apportionment and allocation.

There were still real effects from both rules reform and strategic adjustment, albeit often with unanticipated consequences. Yet they remained reliably insufficient to alter a nominating dynamic forged over time, not to mention the occupational and career seedbeds of those who ran for a presidential nomination within that dynamic and the structural influencesespecially factional preferences and factional struggles-that owned the real selection of a nominee from within that field of contenders. The interaction of strategies and sequences, or the impact of major rules of the game, might have some influence on the result of any given contest, in the sense of making a final idiosyncratic contributions to it. In a sufficiently balanced struggle, this might even help to produce one aspirant rather than another as the nominee.

Yet in the long causal funnel that runs from a historically repetitive bandwagon dynamic to the anointment of one particular presidential nominee, these candidate strategies and state sequences, like formulae for delegate apportionment and delegate allocation, proved to be far down the causal chain. As portentous as one or another of these rules battles may have seemed while it was being waged, or as insightful as a particular strategic gambit may have seemed when it was announced, they could qualify at best-and usually not at all-as final adjustments to the longrunning process that recurrently produced presidential nominees. In that sense, it remained the structural fundamentals that deserved this decisive designation, and they remained largely outside the reach of even the major individual actors.

An argument that the long will-he-or-won't-he ballet was important is Saladin Ambar, American Cicero: Mario Cuomo and the Defense of American Liberalism

By One Vote: The Disputed Presidential Election of 1876

Clay is inevitably also a major connecting thread in Michael F. Holt

Gilded Age Factions

For the larger movement around him, see

For thumbnail accounts of nominating politics and their conventions in what follows, especially when these are used principally in lists of supportive examples, the two simplest references are Bain and Parris, Convention Decisions and Voting Records, along with the CQ Press volume on National Party Conventions

In a very narrow sense, Theodore Roosevelt was "denied renomination" in 1912, though he was actually blocked by the renomination of the sitting president at the time

For a view of the nominating world shortly after their arrival, see Louise Overacker, The Presidential Primary

Primaries

The Making of the President

The Making of the President

Institutionalizing the Disappearance: The Nomination After Reform

The Pure Partisan Institution: National Party Conventions as Research Sites

Media and Momentum: The New Hampshire Primary and Nomination Politics

The Iowa Caucuses and the Presidential Nominating Process

Beyond Bain and Parris plus the CQ Press National Party Conventions, see Michael F. Holt, The Election of 1860: "A Campaign Fraught with Consequences

An Outline of the Public Life and Services of Thomas F. Bayard, Senator of the United States from the State of Delaware

The Presidential Primary (Note 10). The situation at the end of this era is captured in James W. Davis, Presidential Primaries: Road to the White House

Specific results for all direct subsequent references to primary entrants and primary outcomes are drawn from the same compilation

Drawn from Congressional Quarterly's Guide to

For a marvelously rich description of the operation of this system before the plebiscitary era, state by state

The Pure Partisan Institution: National Party Conventions as Research Sites

Party Structures and Representational Impact

Yet 2020 was disturbed powerfully by the coronavirus and its associated pandemic, such that the 2020 process could not always be entered in parallel with preceding years-in which case 2016 became the end of the data series

Texas has provided a sense for the scale of the difference between using the caucus and using the primary, by selecting its delegates in both ways on the same day. See, for example

Jewett refers to them as "the other carve-out states" in Jewitt

The ability of South Carolina to achieve just such an impact for the Democrats in 2020 was to be both anomalous and dependent on an extremely unlikely concatenation of prior circumstances. See the final section of Chapter

Five of the six states created new presidential primaries in 1972 placed them in May, and all five were on days that had not previously hosted a primary in any other state

Busch unpack this southern effort in "The Rise of Front-Loading

The larger struggle over placement within the nominating sequence, and the recurrent role of fights over the "window" within which various contests would occur, is richly considered in Elaine C. Kamarck