The past several decades of elite study have informed researchers that elected officials strive to maintain proximate distance to the policy preferences of their constituents (e.g., Miller and Stokes 1963; Clausen 1973; Kingdon 1973; Fenno 1978; Mann 1978; Herrera and Taylor 1994; Herrera 1995; Sullivan et al. 1993; Stimson, Mackuen, and Erikson 1995; Ansolabehere, Snyder, and Stewart 2001; Canes-Wrone, Brady, and Cogan 2002)--a condition Robert Weissberg (1978) termed "dyadic correspondence." Our political system, however, is teeming with unelected political elites who have enormous policymaking or gatekeeping powers at their disposal (see, e.g., Masket 2009; Cohen et al. 2008), raising the specter that an electoral connection (Mayhew 1974) is a necessary but insufficient condition to bring about what Weissberg further terms "collective correspondence," or the congruence between the aggregated preferences of the public and its elites, systemwide.
However threatening this prospect sounds, several scholars have found that unelected elites can be quite responsive to public opinion. Bureaucrats, for instance, frequently integrate public input into their decision making (Storing 1976; Warwick 1981; Rohr 1986; Burke and Cleary 1989; Kerwin 1999; Goodsell 2004), perhaps even more so than elected legislators (Long 1952), although there is often disagreement about the definitional boundaries of both "responsiveness" and "public wishes" (Saltzstein 1992). Similarly, others (Barnum 1985; Link 1995; Mishler and Sheehan 1996; Flemming and Wood 1997; McGuire and Stimson 2004) suggest that justices on the Supreme Court are sensitive to shifts in public opinion, much as are those who depend on the consent of voters to enjoy the trappings of their office: "Hardly indifferent, these politicians are keen to pick up the faintest signals in their political environment. Like antelope in an open field, they cock their ears and focus their full attention on the slightest sign of danger" (Stimson, Mackuen, and Erikson 1995, 559).
But do all unelected elites exercise such sensitivity to public opinion? One can plausibly argue that judges and bureaucrats are necessarily responsive to the public out of professional necessity; executive bureaucracies are bound by the Administrative Procedures Act and other statutes to consider citizen input throughout the various stages of the decision-making process (Kerwin 1999), while the decisions of federal jurists become part of the public debate, subject to the circumvention and enforcement prerogatives of the legislative and executive branches, respectively. But what about those elites who are not so visible and can exercise a great deal of their power under the cloak of relative anonymity? Are they as responsive to public wishes? This paper focuses on a particular class of elite that received a great deal of attention in the run-up to the 2008 presidential election--Democratic Party Leaders and Elected Officials (PLEOs), more commonly known as "superdelegates."
Since their inception resulting from a 1982 rule change, these unpledged, ex-officio delegates have attracted little attention from scholars who study presidential nominations. This has not been without good reason--most of the Democratic members of Congress, sitting governors, former presidents and vice presidents, members of the Democratic National Committee, and other former and current high-ranking party officials that are superdelegates jump onto the bandwagon of the presumptive party nominee well in advance of the convention, making them inconsequential to the outcome. Not since establishment candidate Walter Mondale bested party insurgent Gary Hart in 1984 with the help of superdelegates has this class of delegate provided any meaningful contribution to the selection of a nominee for the Democratic Party, and even this is subject to debate (Mayer 2009).
Yet in spite of their unromantic past, superdelegates are elites with incredible power at their disposal, should certain circumstances obtain. Such gatekeeping potential was evident in 2008, when a freshman senator barely two years into his term toppled a titan of Democratic Party politics, former first lady and New York senator Hillary Clinton, with superdelegates providing the votes to seal his nomination. Superdelegates were on center stage again in 2016. After strong showings in Iowa and Nevada as well as a solid win in New Hampshire, the outsider candidate and junior senator from Vermont, Bernie Sanders, was virtually tied in the pledged delegate count with Clinton, again the party's frontrunner. Indeed Sanders was within a couple hundred delegates of Clinton as late as March and achieving notable upsets, as in Michigan. However, few astute political observers ever gave him close to even odds in the race since Clinton had chalked up an early and substantial lead in superdelegates. The early disparity in superdelegates despite the electoral context drove almost 200,000 of his supporters to sign a petition calling on superdelegates to follow the will of the people and MoveOn.org pledged to target superdelegates that did not do so (Dickinson 2016).
In this paper, we investigate whether Democratic superdelegates condition their support for a candidate on the expressed preferences of rank-and-file party members in the electorate during a contested nomination, or whether they follow their own preferences. Simply put, do superdelegates face electoral constraint, even if they do not hold public office?
The decision-making process of superdelegates has serious ramifications from the standpoints of democratic theory and party organization. In terms of the former, if superdelegates do indeed consider the expressed preferences of primary voters and caucus-goers, then fears of unelected powerbrokers cutting deals in "smoke-filled rooms" are grossly misplaced. Such was the concern of reformers after the 1968 election, when Hubert Humphrey was nominated by Democratic delegates without competing in a single primary, prompting the creation of the McGovern-Fraser Commission, which ultimately ushered in the "reform" period of presidential nominations. From the view of party organizations, however, such congruence with the party in the electorate stands in contrast to the express purpose of superdelegates, as envisioned by the Hunt Commission that created them: to provide input from party leaders and officials serving in government to moderate the more immediate whims of the party in the electorate. Doing so, it was thought, might prevent unpopular candidacies, such as George McGovern's in 1972, and presidencies, like Carter's (Aldrich 2009). (1)
This tension is manifest in the attitudinal differences between superdelegates and rank-and-file Democratic primary voters and caucus-goers. The latter are far more likely to view the Democratic nomination process as elite-driven and inherently undemocratic (Fridkin, Kenney, and Gershon 2011). Superdelegates, on the other hand, view their ex-officio roles more positively, as serving the best interests of the party and ultimately conforming to popular will, even if they are no more able to assess the ideological location and issue positions of rank-and-file partisans in the electorate than are pledged delegates, who are presumably closer to the people (Herrera 1994). The endorsements from superdelegates who are members of Congress, for example, become more heavily associated with national poll standing over local opinion as the campaign wears on (Hasecke, Meinke, and Scott 2013). Yet, superdelegates have also been shown to vote in line with their own preferences and maintain those to the convention floor (Southwell 2010). Beyond these insights, existing research offers little guidance by way of what to expect of superdelegates' concerns. It is the aim of this study then to establish the degree to which populist considerations influence the endorsement choices of Democratic superdelegates.
Superdelegates' Electoral Connection
The central hypothesis of this study is that superdelegates are sensitive to the political environments in which they operate, particularly the electoral desires of their fellow party members, even if they do not hold public office. We refer to this relationship as the Electoral Connection Hypothesis and leverage the unique nature of the 2008 nomination race to explore this hypothesis in the context of two superdelegate behaviors: (1) the timing of an endorsement and (2) whether or not to switch support.
While the final nomination vote is fascinating in its own right, the decision at the convention is largely a fait accompli in modern nomination campaigns, at least from the perspective of constituent concern, because it is primarily based on the tally of results from the primaries and caucuses. Indeed, many superdelegates often fall-in-line with the candidate who has the most delegates at the nomination in order to publicly signal a unified party and bolster the Party's chances in the general election. Thus there are two opportunities in which superdelegates may be seen to weigh constituent concerns relative to their own and both of these occur substantially prior to the nominating convention: deciding when, if ever, to publicly endorse a candidate, and deciding to switch support to a different candidate. Both of these opportunities were fully available and on display in the ultracompetitive 2008 nomination contest.
The election of 2008 was atypical of most nomination contests. For the first time in over 50 years, neither party was running a candidate as an incumbent, prompting a flood of entrants from each party into the race. On the Democratic side, Clinton early positioned herself as the frontrunner and seemingly inevitable nominee, raising a great deal of money and dominating in the polls (Burden 2010). Before the first caucus vote in Iowa was cast, a New York Times/CBS News poll of 588 superdelegates found 27% of...