Elections, Public Opinion, & Voting Behavior

Every March citizens around the country mark their choices and lay down a dollar in the NCAA basketball tournament office pool. In political science departments, the madness is more likely to hit in November, when many of us try to predict the outcomes of key election contests. With the House and Senate nearly evenly divided between Democrats and Republicans, many of you may be wondering how to wager on the question of party control in the 108th Congress. As a public service to my colleagues in the profession, let me offer some data that may help you decide where – or whether -- to place your bets.

Since 1946, the president’s party has lost seats in the House of Representatives in every midterm election but one. This historical fact would seem to promise a fruitful election season for the Democrats. However, it bears noting that the out-party’s midterm gain has three times – in 1962, 1984 and 1998 -- fallen short of seven seats, the magic number needed to take the majority for the Democrats. If 2002 is a typical midterm election, this suggests a 23% chance (3/13) that the Republicans will keep control of the House. On the Senate side the president’s party has fared a little better at the midterm, actually gaining seats in three of the 13 elections since 1946 (but, incidentally, not the same three elections in which House loss was fewer than seven seats). Still, history appears to suggest that the Senate Democrats will retain the majority in the 108th Congress.

Of course, 2002 is in many ways atypical. It follows a presidential election in which the winner did not have coattails – Republicans lost a net of two House seats in George W. Bush’s slim victory. As Jacobson and Kernell note, “the greater the victorious party’s surge two years earlier, the greater its decline at the midterm” (1983, 61). As a scatterplot (Figure 1) illustrates, this is generally the case albeit with a few very exceptional exceptions – including the post-Watergate election and the Gingrich Revolution -- in which the president’s party was absolutely pummeled in the congressional election. In any event, there is little reason to think that there are significant numbers of Republicans who were swept into the House by a Bush surge, and who are therefore particularly vulnerable in 2002. (In a post-redistricting election, new district lines can potentially disrupt the surge-and-decline pattern. As we shall soon see, this is unlikely to be the case in 2002.) Indeed, political observer Charlie Cook has only classified seven Republican-held seats as “toss-ups” or “leaning Democratic.” This bodes ill for a Democratic takeover: as Cook notes, “if the election were held today, Democrats could win every seat that [is] leaning, likely or solidly Democratic, plus every toss-up race, and still come up two seats short of a majority” (Cook 2002).

None of the Republican Senators up in 2002 were elected with Bush in 2000. For the Senate, the surge-and-decline thesis suggests one look back six years. If national tides sweep one party’s Senate candidates into office, one should expect some of those candidates to lose six years later when the political context has changed. For example, 12 Republicans (nine challengers and three open seat candidates) won previously Democratic Senate seats in the 1980 Reagan landslide. Six years later, half of those freshmen lost re-election bids in a “market correction.” Indeed, in the last 21 elections there is a strong relationship between seats captured, then lost six years later (see Figure 2). Since Democrats actually lost two Senate seats in 1996, their freshmen seem unlikely to face more danger than the Republicans. In fact, two freshmen from each party are listed as “toss-ups” by the Cook Political Report.

Most prognosticators deem both the House and Senate majorities “too close to call,” which should make for an exciting evening of CNN and web monitoring for political junkies. Ironically, the uncertainty about the aggregate results is accompanied by an unprecedented level of certainty about individual matchups. This is all the more surprising given that the 2002 election follows the decennial reapportionment and redistricting. The reallocation of congressional seats and redrawing of district lines has the potential to shake up the electoral landscape by creating new or radically reconfigured districts in which neither the Democrats nor the Republicans field a safe incumbent. Since 1972, the incumbent re-election rate has averaged 90.5% in post-redistricting cycles and 94.7% otherwise; total turnover from defeats and retirements averaged 20% after redistricting and 13% in other years.

In the post-2000 redistricting, the primary consideration in most redrawn maps -- whether created by Democrats, Republicans, both or neither -- was incumbent protection (Mitchell 2002). The effect of this strategy is likely to be a remarkably low turnover for the 108th Congress, as already suggested by the numbers of retirements, primary defeats and uncontested seats (see Table 1). Only 34 House members are not seeking re-election in 2002, compared to 65 in 1992.[1] Only 2.0% of House incumbents lost primaries in 2002, compared to 5.1% in 1992. And 46 incumbents – 11.7% of those participating in a general election – have no major-party opposition, compared to eight (or 2.3%) in 1992. Among those incumbents who do face opposition, only four are in serious danger of losing in November and another 18 may have to hustle to hold onto their seats (Cook 2002).[2] As is customary, Senate elections promise to be more competitive. There are nearly as many fierce contests for Senate as there are for House! Of the 34 races at issue, about eight are currently considered tossups and another seven are being closely watched.

The absence of competition in the upcoming House elections is troubling, no matter what one’s partisan persuasion. Competition gives meaning to the vote choice, for without it the tally can hardly be construed as a substantive expression of citizens’ political preferences. In those few elections where substantive issues may come into play, what are those issues likely to be? Here, too, I hesitate to make a prediction. With only four weeks to go it is still unclear whether any one issue will dominate the discourse. This may be surprising, given the significance and salience of matters on the agenda today. It’s war, stupid! It’s the economy, again, too! How about corporate accountability? Or for the senior voters – a key bloc in many states and districts -- prescription drug benefits. None of these issues (amazingly) is catching fire. Rather, the themes of candidates’ – and groups’ and parties’ -- television ads are incredibly varied (see scripts and video clips in the National Journal’s “Ad Spotlight,” www.nationaljournal.com). Many of them emphasize values and character instead of specific issues, and those perennial favorites, taxes and social security, show up in a number of contests. Health care, gay rights, pollution and immigration are also featured.

It is possible that national themes may emerge in the final month of campaigning. The Bipartisan Campaign Reform Act takes effect on Wednesday, November 6, which means that the national parties must “use it or lose it” when it comes to soft money. The parties have already raised more soft money for the 2002 cycle than they did for 2000; without a presidential election this promises an unprecedented level of spending on party-building and so-called “issue advocacy” (ads that promote a candidate without expressly urging his election). Such a large role for the party committees may introduce some uniformity among Democrats and Republicans, respectively. However, this would break from the patterns established to date. Party-sponsored ads have thus far varied as much as candidate-sponsored ads (Rodeffer 2002).

Cook, Charlie. 2002. “If the Election Were Today.” National Journal.com. September 17. Available http://nationaljournal.com/members/buzz/2002/races/091702.htm. Accessed September 27, 2002.

Federal Election Commission, 2002. “National Party Non-federal Activity Through June 30 of the Election Year.” Web document available at http://www2.fec.gov/press/20020919partyfund/nonfed_history.html. Accessed September 27, 2002.

Mitchell, Alison. 2002. “Redistricting 2002 Produces No Great Shake-Ups.” The New York Times. March 13 (late edition). Page A20.

Stevens, Allison. 2002. “Number of uncontested races extremely high.” The Hill. April 24. Available at http://www2.hillnews.com/042404/c2k2_uncontests.shtm. Accessed September 27, 2002.

[1] Granted, the number of retirements in 1992 was likely inflated by the check-bouncing scandal, but 40 incumbents retired in 1972 and 1982.

[2] The most recent analysis from Cook Political Report, which examines individual House contests, also identifies 18 competitive open seats, for a total of 40 competitive House races. The same analysts found three times as many competitive contests in 1992.