Article | Winter 2007-2008
The Campaign Will Go On!
by Jason Bello (Political Science/Economics/Linguistics, Columbia College) and Robert Shapiro (Political Science)
Left to right: Robert Shapiro and Jason Bello
When we first put together our predictions for the primary election a few weeks before the Iowa caucuses, we sought to look at how the rules of the game constrain the outcome. We noticed that it had been 80 years since there had been an election with no incumbent or early frontrunner in either major party and that the last primary in which the candidates were not decided in advance of the convention was 1976 (Republican). Even more striking, the last convention where there was more than one ballot taken before a winner was declared was 1952's Democratic convention. We asked whether there was anything in the rules governing delegate selection that would lead to an early consolidation of candidates.
From what we found, we argued that after February 5th's super-primary, where 1,191 Republican delegates and 1,585 Democratic delegates will be selected, we will likely still be far from having actual presidential nominees. This only occurs when the candidates garner a majority of the delegates, or approximately 1,200 Republican delegates and 2,200 Democratic delegates. Today (as of January 11, 2008), we stand by this assertion. If anything, the races have become more competitive. It is possible that either party or both could get to its convention with the nominee undecided. This possibility stands in contrast to recent history. The time between one primary and the next is shrinking. The length of the primary campaign has continued to grow. Campaigns have become more professional, and the ways in which the media cover the races have changed. The 2008 election may be a critical test of whether it is reasonably possible for the races to go to the convention undecided. If it does not go to the convention undecided, this race becomes an opportunity to investigate why the conventions have become obsolete in this respect.
The first step is to ask whether or not the institutions surrounding the primaries and caucuses and the ways they distribute delegates facilitate early wins. To answer this, we need to look not only to the news-grabbing metrics like polling, primary dates, and endorsements, but also to the often overlooked rules and regulations of the electoral process. The candidates are ultimately decided not by popular vote, but by delegates at the party conventions. And not all delegates are awarded on the basis of popular vote. In fact, several hundred delegates need only decide their presidential preference at the time of the convention. To make things more complex, the ways in which popularly elected delegates are distributed to candidates vary from state to state: There are proportional schemes, winner-take-all schemes, and several different ways of combining the two.
Focusing on the delegates in this way, we ask a number of important questions. What are the implications of the different distribution schemes? How important are the unbound delegates? Finally, we need to better understand the likelihood of getting to the conventions with no candidate having commitments from a majority of delegates. To answer these questions, we telephoned representatives from the state parties in each continental state and collected nearly all of the hundred delegate selection plans. Simply obtaining all of the information was difficult, as no public disclosure is required, and many states do not post the plans online. We then classified the many systems into a set of types for examining delegate voting.
The whole enterprise of conventions driven by delegates bound by the popular primary vote is a relatively recent invention and a sharp contrast to the back-room politics that dominated the bulk of American history. The evolution of the two major parties, driven by a number of processes including anti-corruption efforts and the opening of parties to broaden representation, has gone in very different directions. The Democratic Party now has a highly centralized system with largely homogeneous plans from one state to another. Some states have primaries, other states have caucuses, but each state chooses the same proportion of bound delegates as unbound delegates (80% bound to 20% unbound) and distributes the bound delegates in exactly the same way-proportionally to every candidate who receives over 15 percent of the vote. The Democrats' system can therefore sustain several candidates-certainly three or four-well into the contest. To get to the convention with the nominee undecided, the difference between the delegate counts of the top two candidates has to be less than 20 percent. The more candidates who can win more than 15 percent of the vote in several states, the more likely it is for this to happen. Based on polls at this point, it seems unlikely that John Edwards will stick around much past February 5th, which decreases the chances of the contest going to the convention undecided. Nonetheless, both of the two remaining candidates, Hillary Clinton and Barack Obama, have strong bases of support, suggesting that there may be a small enough margin just between these two. If the election goes to the convention undecided, Hillary Clinton could have an advantage because she polls well among loyal Democrats (despite some recent big endorsements for Obama), and most of the unpledged delegates are selected by virtue of their party involvement. That being said, these super-delegates are in no way bound to any candidate and might side with Obama if they perceive him as a stronger candidate for the general election.
The Republican Party delegate selection system is entirely different. The Republicans have a highly decentralized delegate selection system. The central party determines the number of delegates each state gets and has some oversight over when the delegates are selected, but the selection methods are left up to the states themselves. This has led to a plethora of selection plans, from New Hampshire's proportional distribution method, which looks very similar to that of the Democrats, to New York's 87 bound delegates all going to the frontrunner. Some states have introduced quirky selection methods. Arkansas gives one delegate to anyone who gets over 10 percent of the vote, gives the rest either to the candidate who gets over 50 percent of the vote (if there is one), or splits the rest among the frontrunners if there is no majority winner. Surprisingly, the Republican delegates, in the aggregate, are also split roughly 80 percent bound to 20 percent unbound, although the unbound delegates are held disproportionately by a handful of states. Because many of the states after Iowa and New Hampshire are winner-take-all, the Republican primary should consolidate quickly after February 5th. The caveat to this is if there is strong regional favoritism, which would lead to a winner in each region and a large number of national candidates.
How do all these plans of state parties come together in terms of what we know about the upcoming election? As a general rule, the Democratic primaries more or less mirror public opinion. This fact makes the Democratic primaries very susceptible to swings in momentum, where success in one state helps a candidate in later contests. The Republican side is more complex. Imagine for simplicity that the value of winning a state is a function of the number of delegates up for grabs, and the momentum effect it has on later states. For Democrats, the first is more or less constant because the bulk of delegates are distributed on the basis of population and awarded proportionally. Therefore, competing in the early states has a decided and uncompromised advantage. For the Republicans, winner-take-all states are much more valuable than proportional states, and these all come on or after February 5th. The "late state strategy" that Rudy Giuliani has taken is therefore feasible on the Republican side if the candidate believes that the gain from big wins later on is greater than and is uncompromised by negative momentum from early losses.
All of the early states with local favorites (such as Utah for Mitt Romney, New York and New Jersey for Giuliani, Arkansas for Mike Huckabee and Arizona for John McCain) have some form of a winner-take-all delegate distribution system. Giuliani is fortunate in that of all the candidates, his core states in particular amount to a sizeable chunk of the available delegates. In New York, New Jersey, Florida, and California alone, Giuliani could win as many as 250 delegates. Early on we argued that Florida was a key state to watch. Before the February 5th multi-primary day, Florida's primary may reinforce Giuliani's viability to undecided voters if he can win. Huckabee has recently turned Florida, formerly a Giuliani stronghold, into a competitive state. This primary will illuminate whether the negative momentum, so to speak, from the earliest states is strong enough to destabilize Giuliani's late state strategy. For its symbolic value alone, Florida is important to watch. Florida also has a sizeable delegation; even though the Republican Party has penalized it for holding its primary so early, it still has 57 delegates, and all of them will go to the winner.
There is another critical piece of the primary rules beyond simply the delegate counts-whether or not Independents are permitted to participate in the primaries. This rule has received considerable attention because both parties have candidates who are favored by Independent voters-Obama (over Clinton) on the Democratic side and Giuliani and McCain (over Romney and Huckabee) on the Republican side. All the contests so far, Iowa and New Hampshire, have been functionally open, meaning that anyone can vote in any primary or caucus. In the close to forty contests remaining, two-fifths are closed contests, and the remaining three-fifths are open to Independents (several of these are open to members of another party). Some states like Iowa have nominally closed contests that are circumvented by allowing same-day party registration switches.
What does this mean for the candidates? For one, it means that the favorites of Independents should compete vigorously in open contests, and party insider candidates should compete most strongly for the closed contests. This illustrates a critical mistake in Giuliani's late state strategy. By not competing in New Hampshire, an open primary state with a lot of Independents and high voter turnout, he forfeited his moderate Independent votes to McCain, who is, in reality, quite conservative. It also suggests that Giuliani (at the time of this writing) may have made a mistake by not seriously competing in Michigan, another open contest, which could have been a showcase for his support.
Finally, we can connect the rules concerning Independents to momentum. Scholars talk of momentum in terms of perceptions of viability, individual utility of voting for a winner, and social cue-taking. They also pay attention to expectations. It is not winning or losing that produces momentum, it is doing better or worse than expected. Bill Clinton's big win in New Hampshire in 1992 was a second place finish. What happens when we combine this interpretation of momentum with the rules? Expectations are set without regard for the rules, but the results are very much affected by the rules. In closed primary states, there is a decided advantage for party insider candidates like Clinton. However, the expectations will be set largely on the basis of the states where Independents are allowed to participate. This could bring about a de facto advantage for insider candidates who do better in part because of a difference in rules.
Just out of New Hampshire, with at least four viable candidates in one race and at least two in the other, this year's presidential nominee contests have all the conditions for lasting until the convention. We find no reason that the delegate selection process facilitates an early win; on the contrary, we think it might make things more competitive. If these nomination processes end before the convention, we will need to take a step back and ask whether or not the convention has continued relevance as a means for selecting the nominees, and what factors (including especially the strength of early momentum) have caused its decline.
Jason Bello, a former ISERP Undergraduate Fellow, is majoring in Political Science and Economics and minoring in Linguistics at Columbia University. Bello is a 2008 recipient of the prestigious Rhodes Scholarship and plans to pursue a master's in comparative government at Oxford in the fall.
Robert Y. Shapiro is Professor of Political Science, Director of the Public Opinion Project at ISERP, and President of the New York Chapter of the American Association for Public Opinion Research (NYAAPOR). He specializes in American politics with research and teaching interests in public opinion, policymaking, political leadership, the mass media, and the applications of statistical methods.
