Political Cartograhy: The Emergence of the New Incumbent Safety in the U.S. House

By Richard Born

More than two decades ago, a prominent student of congressional politics, Charles O. Jones, made the following observation: "I am convinced that one more article demonstrating that House incumbents tend to win re-election will induce a spontaneous primal scream among all congressional scholars across the nation."

Why, then, am I asking Vassar Quarterly readers to bear with me on the subject of incumbent safety and risk damage to their own vocal chords? The short answer is that the sources of electoral security have undergone a striking transformation in recent years, and this new form of safety is creating a new legislative body. A "permanent," or at least semi-permanent, Republican House majority may have come into being. But unlike the earlier Democratic majorities of the 40 years prior to 1995, the new Republican regime seems destined to be one of razor-thin majorities and razor-sharp partisanship.

The old electoral order was one ruled by Speaker Tip O'Neill's (D., Mass.) famous maxim, "All politics are local." The secret of congressional longevity in good part rested on performing a combination of service activities (e.g., tracking down lost Social Security checks) and pork barrel activities (e.g., earmarking appropriations for federal office construction in the district). Such activities were "pure profit," rallying both Democratic and Republican constituents behind the member in a cross-party coalition. Typically in the 1980s, almost half of identifiers with the challenger's party would be induced to cast a pro-incumbent ballot.

While different studies disagree as to the precise timing of change, all agree that by the early 1990s, incumbency status per se was no longer so much an entitlement to re-election. My own research shows that in a model simultaneously accounting for both House and presidential on-year voting in terms of voters' issue preferences, partisanship, economic evaluations, assessments of the presidential candidates' personal qualities, and demographic characteristics, the electoral value of being an incumbent rather than an open-seat candidate fell 16 percent, on average, from 1980-88 to 1992-2000. An analogous model of midterm voting, necessarily absent the presidential voting equation and the presidential candidate variables, reveals comparable decline in the power of incumbency from 1978-86 to 1990-98.

Across the same expanse of time, however, House elections have become more nationalized in the sense that voting is now more determined by voters' partisanship, as well as by their ideology and positions on enduring issue concerns, like abortion rights, that themselves have come more into line with partisanship. Thus, the gradual substitution of party- and ideology/issue-based cues for incumbency cues may have led to the decline of safety. Alternative explanations include the possibility that voters increasingly have been taking for granted the constituency solicitude now automatically associated with the job of Congress member, and greater national party efforts to recruit and finance capable challengers to incumbents. Whatever the responsible cause or causes, incumbents' average vote-share during the first four elections of the 1990s slumped to its lowest level since the decade of the 1960s.

Losing some control over their own fate was certainly discomfiting to the famously risk-averse members of Congress, who typically view election safety the way Babe Paley regarded richness and thinness. But from the collective standpoint of the Democratic congressional party after 1994, greater incumbent vulnerability translated into the possibility that a modest anti-GOP swing nationally could topple enough Republicans to transfer party control. Subsequent events relating to congressional redistricting at the start of the most recent decade, however, fly in the face of this possibility, making the House the most difficult part of national government for the Democrats to switch.

Even before the 2001-02 round of redrawing district lines, state legislators in charge of the process commonly engaged in what has been termed bipartisan gerrymandering. Simply put, this means making Republican districts more Republican and Democratic districts more Democratic; since members overwhelmingly seek re-election, the alternate name for such redistricting is pro-incumbent gerrymandering. During previous redistricting cycles, the most important reasons for doing things this way were a desire to avoid the bitter struggles for partisan advantage that could paralyze other state legislative business, lobbying by self-interested members of the U.S. House delegation, and bipartisan compromise forced by divided party control of state government. All three factors operated in 2001-02. An additional development, however, was that even in some states where Democrats controlled redistricting, or where Democrats were tempted to fight it out with the opposition, despite divided party control, these legislators eventually resigned themselves to a bipartisan gerrymander. In the former case — notably, California — Democratic U.S. House members were reluctant to accept the large-scale transfers of constituencies that normally are required for partisan gerrymandering. In the latter case, there was fear that if stalemate with Republicans resulted, the subsequent obligatory intervention by a federal panel of judges might lead to a final redistricting map that was worse for Democrats than the status quo.

State legislative Republicans, by way of contrast, were far more aggressive in their approach to redistricting. In the large GOP-controlled states of Florida, Michigan, Pennsylvania, and Ohio, partisan plans were enacted with the intention of boosting Republican representation by one to three seats each, despite declines in the overall number of districts apportioned to the latter three states caused by below-average population growth.

Still, intentions do not necessarily equate with election day realities — past partisan gerrymanders have sometimes fallen prey to the so-called backfire effect. The classic partisan gerrymander technique involves creating a maximum number of districts with lean majorities for the redistricting party, while packing opposition voters into a minimum number of super-majority districts. What can happen is that the redistricting party cuts things too close in estimating how voters will behave, and its candidates end up losing narrowly.


a drawing of two hands, each pulling at opposite ends of a map
a drawing of two hands, each pulling at opposite ends of a map

For two different reasons, however, such miscalculations are now less likely, and the GOP in 2002 was indeed able to realize all but one of its intended gains in the four states just mentioned. First, redistricting architects have access to much more sophisticated computer software that can break down at the level of very small residential aggregates a wide variety of information about voters' past partisan behavior and demographic characteristics. An almost infinite variety of possible districting schemes then can be churned out and evaluated for likely partisan impact. Second, as already stated, voters have become more regularly partisan in their House voting behavior. So computer forecasts of a Republican electoral majority in a district probably will prove as reliable as predictions that the Cubs will not vanquish the Billy Goat Curse in October 2004, and almost surely will prove more reliable than Dan Rather's early, election-evening 2000 promise about CBS's Bush vs. Gore projections. ("If we call a state, you can take it to the bank. Book it!")

The 2001-02 redistricting cycle, therefore, produced unusually large numbers of very safe districts in states with bipartisan gerrymanders, and moderately safe but dependable districts in states with partisan gerrymanders. This is emphatically borne out by the 2002 House returns. The number of incumbents losing to challengers was only 4 of the 389 attempting re-election, the lowest in U.S. history. (Four others were ousted in incumbent vs. incumbent matchups.) Eighty-three percent of incumbents received at least 60 percent of the vote, compared to only 66 percent who did this well in 1992, the last election that immediately followed redistricting. It used to be pointed out when I was studying incumbents in the 1980s that their chances of being voted out of office were comparable to those of members of the old Supreme Soviet, the sham legislature of the USSR. Today, a more apt comparison might be to the British House of Lords.

And things may soon get worse. Current redistricting practice involves adjusting boundaries only once per decade, just after the release of data from the new Census. The only cases of mid-decade redistricting — at least over the past 11 decades — have been when the federal judiciary finds a plan unconstitutional, generally because of unacceptable population variance among districts, and orders a revised plan implemented. Texas Republicans, however, spurred on by native son and aspiring Speaker Tom DeLay ("I'm the Majority Leader, and we want more seats"), enacted a mid-decade plan on their own initiative in late 2003 that qualifies as the most ambitious exercise in creative political cartography since Kuwait was declared to be the 19th province of Iraq in 1990. (A court-devised plan was adopted in 2002 after opposing party majorities in the two Texas houses deadlocked, but that same election year subsequently gave the Republicans control over both chambers.) The GOP's best-case scenario was to eliminate every one of the 10 Anglo Democrats chosen in 2002, half of whom were narrowly reelected by margins of under 60 percent. Seven, it was hoped, would be replaced by Republicans, and the remaining three by African-American or Hispanic Democrats. Thus, the 17-15 Democratic advantage at the end of 2003 would shift to a 22-10 Republican edge after the 2004 election, with nobody left but safe Republicans and even safer Democratic minorities. Already, the plan has caused Congressman Ralph Hall to switch his party allegiance from Democratic to Republican. In response to Democratic legal challenges, the U.S. Supreme Court in mid-January refused to stay implementation of the plan.

Regardless of how Texas ends up, we seem to have settled, at least until 2012, into an era of a narrowly Republican-controlled House, suffused with unprecedentedly safe incumbents from both parties. The form of that safety, of course, is different than that of the past, less dependent on performing service and pork-barrel functions, and more dependent on gerrymandered partisan majorities and more steadfast party voting by these district partisans. What does this imply for the operation of Congress? Most obviously, the party polarization of members that gathered momentum in the 1990s can be expected to escalate. A mere handful of House Democrats today are more conservative than the most liberal Republican. In the past, members from competitive districts had to create a personal cross-party constituency for themselves not just by providing material benefits, but by throwing roll-call votes in both ideological directions. A world of gerrymandered districts is a world in which smart electoral politics means giving a district's majority partisans what they want by toeing the party line.

Furthermore, the House is the body of Congress where Republican leaders have the stronger policy incentive to stake out immoderate positions on issues, which dovetails with their members' own electoral incentive to support such legislation. In the House, simple, extreme majorities can rule, whereas in the Senate, because extraordinary majorities of at least 60 senators are required to defeat the threat of a filibuster, policy usually must be less extreme. The routine since the Republican congressional takeover in 1995 increasingly has been to take very conservative House, and moderately conservative Senate, bills to Conference Committee and then to adopt a compromise closer to the House version. This often is facilitated by shutting almost all Democrats out of the meaningful Conference deliberations that then lead to a pro forma, party-line final vote. The critical strategic terrain occupied by the House was well appreciated by Republicans like Newt Gingrich when they waged their great battle in 1994 to overturn 40 years of Democratic dominance, and it helps account for the ferocity of their efforts to retain control ever since.

Additional breakdowns in what little remains of Republican-Democratic comity seem inevitable in today's take-no-prisoners House. I have in mind events that might resemble Ways and Means Chair Bill Thomas using the Capitol Police to disband a Democratic strategy session on a pension reform bill last July, or Speaker Dennis Hastert holding open for more than two and a half hours beyond the official 15-minute limit the floor vote on Medicare prescription drug benefits last November, so that enough GOP dissenters could be pressured to switch from nay to aye. Were the Democrats to regain the majority (through virtual reality?), there is no reason to think, given the political realities I have outlined, that cross-party coalition building would become any more in fashion. But while the House drifts away from moderation, the U.S. electorate-at-large remains almost precisely balanced between Republican and Democratic identification, and this "perfect tie" is unlikely to be broken any time soon. So the combination of extreme policies and the hardball tactics required to enact them may prove very alienating to voters, causing us to long for the days when party elites offered more echoes and fewer choices.

Suggested Reading

Brady, David W., Robert D'Onofrio, and Morris P. Fiorina. 2000. "The Nationalization of Electoral Forces Revisited." In Continuity and Change in House Elections, ed. David W. Brady, John F. Cogan, and Morris P. Fiorina. Stanford, CA: Stanford University Press.

Cohen, Richard E. 2003. "Broken Barometer." National Journal, 12 July, 2003, 2240-2249.

Fiorina, Morris P. 2001. "Keystone Revisited." In Congress Reconsidered. 7th edition, ed. Lawrence C. Dodd and Bruce I. Oppenheimer. Washington, DC: CQ Press.

Jacobson, Gary C. 2003. "Terror, Terrain, and Turnout: Explaining the 2002 Midterm Elections." Political Science Quarterly 118: 1-22.

Toobin, Jeffrey. 2003. "Annals of Law: The Great Election Grab." New Yorker, 8 December, 2003, 63-80.

— Richard Born

Richard Born, professor of political science, teaches the following courses: "American Politics," "Congress," "The American Presidency," "Political Parties and Public Opinion," "Political Analysis," and "Seminar in Congressional Politics." His research interests are in the areas of congressional and presidential elections. Most recently, he has completed a study on the nationalization of voting for the U.S. House from 1978 to 2000.