Incumbency and the Personal Vote

Russell D. Renka
Department of Political Science
8 February 2000

°PS365 Syllabus
°Renka's Home Page
°Web Links

    The dominant element of congressional elections is the difficulty in defeating incumbents.  This is especially true in the "lower house" with its short election terms of two years, designed in the founding to ensure that one house of one branch would be closely responsive to popular movements with their inherent electoral jeopardy.  This brief paper demonstrates the links between modern electoral safety, careerism, and the personal vote in a highly institutionalized Congress that has moved far from "original intention."  I forecast that a result of continuation of recent trends will be national adoption of a congressional term limitations amendment to the U.S. Constitution.

The Evidence

    From 1946 through 1998, a period of 52 postwar years with 26 elections divided between 24 midterms and 22 presidential years, the U.S. House of Representatives incumbents who ran for another term of office were successful 91.8 percent of the time (Davidson and Oleszek 2000, Table 3-1, pp. 65-66 as corrected per Ornstein et al., Table 2-7, p. 57).  Senators won 78.2 percent during the same time span.  For the 425 members of the House, an average of 396.5 chose to run for reelection, leaving an average retirement rate at 40 Members per election of just under 10 percent.  Those 397 (rounded off) saw an average of 7 losers in state partisan primaries and another 26 in the general election.  So 33 departed the electoral way, but more than that (40) declined to run again.  The average turnover of 73 Members represents approximately an 18 percent turnover rate for each new Congress.

    The evidence since the 1980 census is even more strikingly favorable to incumbents.  The nine elections in the 1980 and 1990 census cycle from 1982 through 1998 shows House incumbents winning at a 94.3 percent rate despite two elections (in 1992 and 1994) that both produced substantial upheavals reflecting public antagonism toward status quo congressional politics.  Senators raised their rate to 88.6 percent, a full ten percentage points higher than the 78 percent long term rate.

    Now let's consider the meaning of these success rates for careerist Members and for those who would be, namely serious political aspirants who carefully judge their chances of replacing an incumbent.  A benchmark for congressional incumbent safety is a 90 percent reelection rate, and that applies as well to many state legislatures.  The House incumbent success rate from 1946 through 1980, for instance, was 90.2 percent.  Now consider what this means over time.  Each biennial election promises a 10 percent chance that one's career ends.  How likely is it, then, that a newly elected member would fail to serve 12 years?  I assume for simplicity here that death and disability do not intervene, so the only pitfall is electoral defeat.  I do not consider personal characteristics or district variation, so the median member is the benchmark.  I also use the 12-year benchmark because that is the prevalent proposal of serious term limitation advocates.  The answer is obtained simply by taking 10 percent raised to the power of (number of elections), namely five; so for any x terms, the exponent for new incumbent's likelihood of defeat is (x-1).  A twelve-year aspirant therefore has a 41 percent (.40951) likelihood of losing.  Ambitious challengers and the opposition party will know this.

    For senators, only one reelection is required to achieve 12 years.  Incumbents from 1946 through 1980 won at a 72.3 percent rate.  If the incumbent success rate is taken as 75 percent, then a .25 likelihood of losing is known to senatorial aspirants (including the numerous House members who deem themselves worthy of sitting in the "other body").

    Now instead of 90 percent success rate for House incumbents, use 94.3 percent to represent post-1980 experience.  Now a 12-year aspirant has only a 23.4 percent (.2343) likelihood of losing; that's barely over half the figure for 90 percent success per election.  And a Senator at 88.6 percent success has just over 11 percent chance of failure, less than half the former chance.  In 1998 there were new extremes, with House incumbents chalking up 394 wins and just 7 losses, a 98.3 percent success rate helped along by a record 93 incumbents facing no opponent in November.  Every percentage point above 90 is a powerful deterrent to challengers in biennial House races, and even Senators look out of reach when an 11 percent chance exists only once in six years.  Congress takes on the appearance of a closed shop.

The Term Limitations Movement

    The Congress does not enjoy high esteem from the American people.  Term limitations took off beginning in 1990, producing more than two dozen applications to state legislatures by an average two-to-one popular vote margin.  The movement is closely associated with the dismally low approval ratings the public conferred upon Congress in an era of chronic bipartisan impasse over the huge budget deficits run up by the federal government in the 1980s.  Gallup Poll Trends - Congress Job Approval shows that in 1990 the institution won only a 25 percent approval rating against a 65 percent average disapproval.  The nadir was 1992, with a single yearly rating in March 1992 of 18 percent approval and a lofty 78 percent disapproval.  Incidentally, that is the month of the Illinois Democratic primary in which an incumbent Senator was soundly defeated by an unknown Chicago challenger, owing primarily to the fact that that month was the exact locus of the House "banking scandal."  The fact that Senators had a separate bank and no involvement in the scandal did not prevent voters from issuing summary judgment against the institution and any who happened to represent or personify it during that unfortunate month.

    State legislative term limitations were adopted in three states in 1990, another 10 in 1992, and five since then (Washingtonpost.com Term Limits Special Report) for a tally of 18.  Half the states adopted an 8-year limit on state lower house service, and 13 applied the same limit to the state senate.  Many of these terminations for those elected in 1992 take effect in 2000, leaving those incumbents either to run for the other body, aim for a federal office, or turn their ambitions elsewhere.  

References

Davidson, Roger H, and Walter J. Oleszek.  2000.  Congress and Its Members, 7e.  Washington, D.C.:  Congressional Quarterly.