The extensive literature on the decline of political parties has described the significant changes in party organizations and their relevance to the nomination of presidential candidates. What this literature has so far neglected is the explicit link between nomination procedures and the possibility of maintaining democratic governance. Various authors have speculated that the evolution of the presidential nominating procedures from a party-brokered system to a candidate-centered system should have profound implications for the way in which candidates act during the campaign and once they assume office,(1) but few studies have explicitly explored the probable links between nominating procedures and governing activities.(2)
In this paper, we argue that as presidential campaigns have evolved from a brokered-convention system to a candidate-centered system, we should expect to see a shift in the way in which presidential candidates relate to the various elements in their nominating, electoral, and governing coalitions. We first look at the number of presidential issue stances taken by the candidates. We then compare these positions to the party platforms. The rate of agreement between the presidential candidates and their parties' platforms indicates the extent to which candidates take seriously the various elements of the coalitions that nominated them. Second, we categorize the types of issues that presidential candidates address. The types of issues are relevant both to campaigns and to governance. The types of issues reveal the coalitions, appeals, and strategies on which candidates will rely If the actors in presidential campaigns have changed over time, then we should expect changes in the nature of the appeals candidates make. Finally, we compare the stances of the candidates to positions taken by congressional leaders. If congressional leaders have truly been excluded from the process of nominating presidential candidates, then we should expect candidates to articulate a greater number of positions that diverge from those taken by the party leaders. Congressional leaders once exercised a modicum of control over the presidential selection process. In the absence of such control, congressional leaders and presidential candidates are less likely to agree.
Data and Methods
Presidential issue stances are taken from a content analysis of the New York Times. These stories include transcripts of speeches, transcripts of news conferences, reports of speeches by candidates, transcripts of press releases from campaigns, and interviews of candidates by reporters. We analyzed all stories in the New York Times related to the campaign activities and pronouncements of the candidates. The recording unit is the sentence, and the context unit is the paragraph. We read and coded all sentences in the stories and transcripts for issue pronouncements. We followed a procedure developed by Pomper and Lederman that categorizes policy pronouncements according to the specificity of the policy.(3) We included all stories and transcripts from the New York Times from the end of the nominating convention to election day.
We applied the same coding rules that we used for the presidential candidates' positions, taken from the New York Times, to the coding of issue stances for party leaders from Congress. However, we took issue stances for the party leaders from both the New York Times and the Congressional Quarterly Weekly Report and extended the time frame from which we coded their positions. We expanded the base for party leaders for two reasons. First, the New York Times does not substantially cover the activities and speeches of congressional leaders. Second, party leaders can be involved in shaping the policy agendas during the presidential nominating campaigns. Including the Congressional Quarterly Weekly Report and expanding the time frame increases the number of policy positions taken by congressional leaders. Consequently, we can compare the positions of the presidential leaders to those of the congressional leaders. Patterson contains a detailed discussion of the coding rules and the list of the congressional leaders.(4)
The placement of issue stances into policy categories is a relatively straightforward task. We used the definitions of policy types and the ensuing categories outlined in Ripley and Franklin.(5) We assigned each presidential issue stance to one of six categories: redistributive, distributive, competitive regulatory, protective regulatory, strategic, or structural. We omitted crisis policies from the analysis.
Presidential Promises and Party Platforms
Party organizations and the partisan orientations of voters still affect elections for lower offices, but the national parties rarely act as intermediaries for the voters and the presidential candidates. Of course, political parties have not been entirely replaced, but the media, interest groups, and individual candidates' organizations now wield much more influence in national campaigns.
The inclusion of these additional players undoubtedly has complicated the task of assembling electoral and governing coalitions from nominating coalitions. But have changes in the system of nominating presidential candidates precluded the formation of overlapping nomination and general election coalitions? Part of the answer to this question can be found by looking at the number of presidential issue stances over time and by assessing the extent to which the party platforms and the presidential candidates' policy positions agree.
As the clout of political parties wanes in a candidate-centered system, presidential candidates must fashion nominating and electoral coalitions with little or no assistance from party organizations. These coalitions can be heterogeneous and may vary substantially from traditional stands of the political parties. Consequently, we can expect that as the candidate-centered system of presidential politics evolves, the number of positions that candidates take would increase. This hypothesis, that presidential candidates make more campaign promises in the postreform era than in the prereform era, is supported by the actual number of promises made by each presidential candidate between 1952 and 1996 (Table 1).
Numbers of Issue Positions Held by Candidates in Presidential Campaigns: 1952 to 1996
Year Republicans Democrats 1952 45 39 1956 34 111 1960 82 80 1964 60 37 1968 76 70 1972 22 95 1976 55 77 1980 89 84 1984 31 48 1988 102 119 1992 85 111 1996 87 86 Total 768 957 Mean 64 80 Standard deviation 26.49 27.60 Pre-1972 mean 59 67 Post-1972 mean 75 89 Presidential candidates seem to make ever more promises with each subsequent election. Michael Dukakis achieved a high of 119 positions in 1988. Democrats make more promises than do Republicans, but that difference can be explained partly by the larger number of incumbent Republicans. Incumbents tend to be stronger candidates. They run on their records of accomplishments and do not need to take large numbers of issue stances to introduce themselves to voters (Table 2). Incumbent presidential candidates are a known quantity. Bill Clinton, widely viewed as a weak candidate in 1992, took a whopping 111 issue positions. What a difference four years makes. In 1996, President Clinton took only 86.
TABLE 2 Average Numbers of Presidential Campaign Issues by Era and Incumbency Incumbents Nonincumbents Prereform era 36 70 Postreform era 61 91 Average for both eras 57 79 The raw number of issue positions does not reveal the entire story of how the ties between elements of the nominating and electoral coalitions might have changed as a result of transformations in the electoral system. If they are to win the presidency, candidates must assemble coalitions to win the nomination. Then they must strengthen, solidify, and broaden these...