Study Shows Incumbents More Likely to be Re-Elected Under “Top-Two”

A look at both congressional elections, and state legislative elections, in the two states that have used “top-two”, shows that when states switch to “top-two”, the ability of incumbents to be re-elected increases. Also, when a state switches away from “top-two”, the ability of incumbents to be re-elected declines.

Washington used a classic open primary in 2006, but switched to a “top-two” system in 2008. No congressional incumbent lost in either election year. For Washington state legislative races, in 2006, six incumbents were defeated (in Senate districts 6 and 44, and House districts 6-1, 23-2, 24-1, and 31-2). But in 2008, under “top-two”, five incumbents were defeated (in Senate district 2, and House districts 6-1, 6-2, 17, and 44). It is true that there isn’t much difference between six incumbents losing under one system, and five incumbents losing under another system, of course. Nevertheless, the fact that more incumbents lost under a non-top-two system than lost under the first “top-two” experience in Washington contradicts the extravagant promises made by supporters of “top-two” in California. These supporters assume that “top-two” will wash away the “extremists” in the legislature.

Louisiana’s experience is more dramatic. The first year Louisiana used “top-two” for its state legislative races, in 1975, only five incumbents were defeated, our of 144 races (Senate district 38, and House districts 45, 61, 99, and 100). Unfortunately I have not been able to get the data for 1971, the last year Louisiana used a closed primary for state legislative races, but I will continue trying.

In Lousiana Congressional races, during the years that state used “top-two”, out of 112 U.S. regularly-scheduled U.S. House races, only one incumbent was ever defeated (except that in 1992, when two incumbents were left without a district due to reapportionment and redistricting, two incumbents had to run against two other incumbents, so obviously two incumbents lost). No incumbent was defeated in the ten U.S. Senate races. But when Louisiana switched back to semi-closed primaries for Congress starting in 2008, two U.S. House incumbents (out of 7 districts) were defeated.


Study Shows Incumbents More Likely to be Re-Elected Under “Top-Two” — 9 Comments

  1. It does not matter if Top Two resulted in the defeat of all incumbents. Every democratic nation in the world has ALL PARTIES on a general election ballot. PERIOD. Anything less is Fascist.

  2. In 2008, one of those Louisiana congressmen was under indictment, and had the FBI seize $90,000 in cold cash from his freezer. While it was true that he won re-election in 2006 under the open primary, he was challenged, and might have finally been beaten in 2008 if the open primary had been retained. In 2008, the general election was delayed until December because of Hurricane Gustav. Turnout in the “general election” was 40% of the turnout of that in the Democratic primary in November, which had been coincident with the general election for US Senator and President.

    In the other seat, the “incumbent” representative had only been in office for 6 months, and had a narrow win in series of 3 special elections. Voters were dissatisfied that the previous 11-term representative had resigned mid-term to become a lobbyist. So it was more a case of weak and vulnerable incumbents, and some special circumstances that knocked off the incumbents, rather than the return to the partisan primary system.

    Your fundamental premise that the number of incumbents that are defeated is a sign of the quality of an electoral system may be flawed as well.

    Consider a State with a high divorce and re-marriage rate, where the incumbent spouse is replaced. One could argue that this is a healthy situation, because married people recognize that their marriage was deeply flawed, and rather than trying to stick it out while being miserable, took decisive action. On the other hand, it may mean that marriages were entered into frivolously and capriciously in the first place.

    If one wanted a system that gave incumbents at least a 50% chance of being re-elected, but maximized incumbents being defeated, the optimal system would be a coin flip.

  3. Equal nominating petitions.

    P.R. and A.V.

    NO primaries, caucuses and conventions are needed.

    Every election is NEW and has ZERO to do with any prior election.

    Due to the current gerrymanders about 97 percent of incumbents win — no matter what the other stuff is in the rest of the election system — plurality nominations, plurality elections, top 2 or whatever.

  4. “[California “top two”] supporters assume that “top-two” will wash away the “extremists” in the legislature.”

    The “top two” proponents never get around to actually naming any of those so-called “extremists.”

    The outcome of Louisiana’s 1991 and 1995 gubernatorial races belies the contention that the “top two” favors “moderates” or “centrists.” The 1991 runoff featured ex-Gov. Edwards (D), who had already been tried for fraud, and David Duke (R), an ex-Grand Dragon of the Ku Klux Klan. The incumbent governor, who had switched to the Republicans in March 1991, finished third; he had the endorsement of President Bush I and the national Republicans. Another Republican was backed by the state GOP, and he ran fourth.

    Duke was repudiated by the Republican leaders, all of whom supported the Democrat Edwards in the runoff (a similar thing had happened in the 1990 U. S. Senate race). Many Louisianans held their noses and voted for Edwards, who won with 61%.

    The 1995 gubernatorial runoff featured a white conservative Republican and a black liberal Democrat; now-US Sen. Mary Landrieu, a moderate Democrat, had finished third, and ex-Gov. Roemer (R) ran fourth. Like the ’91 runoff, the ’95 runoff was not competitive, as the Republican, Mike Foster, won with some 66%.

    The ’91 and ’95 elections are examples of how the “top two” enables extreme candidates like Duke and Cleo Fields (’95) to reach the final election. If Louisiana had had party primaries, neither Duke nor Fields would have been on the general election ballot, as they would have been beaten in the primary (remember that, when LA has party primaries, 50%-plus is required to win).

    In the 1996 race for the open US Senate seat, there were five or six serious Republican candidates, and it seemed likely that they would split the GOP vote and enable two Democrats to make the runoff. At the 11th hour, the GOP leaders endorsed Woody Jenkins, who made the runoff against the Democrat Landrieu.

  5. “… 1971, the last year Louisiana used a closed primary for state legislative races…”

    In 1971, the Republicans were still only running a few candidates; I doubt that there were more than a handful of Republican candidates for the legislature– if that many. Louisiana was still largely a one-party state.

    The one-party system was a de facto “top two” system, except that a voter had to register as a Democrat to participate in the one-party system.

    When the Republicans started running a few candidates, the winner of the Democratic primary faced a third campaign (since a Democratic runoff was usually necessary); his Republican opponent, in contrast, usually only had to campaign in the general election. This was the big reason that Democratic politicians in Louisiana and Mississippi wanted to implement the “top two” (popularly called the “open primary”).

    Many voters also liked the concept of the “open primary,” since they were accustomed to (1) choosing among ALL the candidates in the first round, and (2) having elections decided with no more than two rounds of voting.

    Hence Louisiana’s “open primary” is a remnant of the old one-party (truly no-party) system.

  6. Washington has a legislative districting scheme which is fairly unique, where senators and representatives run from the same district, with voters electing legislators from three separate positions. Rather than voting for two representatives, the representatives are chosen from two separate positions.

    In some cases this has resulted in bi-partisan delegations. Even though the electorate was identical, voters might elect Bob the Democrat to one position, and Sue the Republican to the other. In districts that are strongly partisan, this of course does not happen. If a district is 70% Democrat it will have 3 Democratic legislators, and if is 70% Republican it will have 3 Republican legislators.

    But in areas that have a more balanced partisan split, there may be a split delegation. A certain percentage of the population will vote for Bob the former city councilman and Sue the former school board member. And some voters will deliberately pick a member from both parties. In choosing the Democrat, they might pick a long term incumbent, even if the district is trending Republican, and perhaps vote Republican for an open seat.

    If there are two Democratic incumbent representatives, there is probably one that is perceived as being a better representative, or at least stronger politically. Since they are elected by position, a stronger Republican challenger may file against the weaker candidate – or if a position is an open seat, file for that seat. So you may produce a split delegation that way.

    You may also have a party split between the senator and the two representatives. Because senators are elected for four year terms at alternate elections, they may miss re-aligning elections where Democratic legislators are beaten in an election along with Democratic federal legislators and statewide officials (Washington elects governors and most statewide officials in presidential election years so legislative races may decided by coat tails in certain years and not other). You might have both Democratic representatives in a district swept out of office in one election, and the Democratic senator re-elected two years later. Senators will tend to have greater length of service since some will have first been elected to the House before “moving up” to the Senate seat. So voters will be more familiar with them as individuals.

    The blanket primary may have supported this bi-partisanship. If a moderate Republican can attract some independent and Democratic support, they may be able to defeat a more partisan Republican in a blanket primary. They may also block emergence of a strong Democratic challenger. If some Democrats and independent voters are voting for a Republican in the primary, the electorate for the Democratic nominatin will be more partisan.

    Over the long term the number of bi-partisan delegations had been declining. This may be due the increase of mass communication drowning out more local communication. Voters in large cities are more likely to simply to vote (R) or (D) due to lack of other information.

    The two years under the Pick-A-Party primary increased this trend as legislative nominees became the choice of partisan voters, rather than the whole community. The incumbents that were beaten tended to be from mixed delegations, which now became 3-0 for one or the other party.

    The introduction of the Top 2 primary should not be expected to instantly reverse this trend. The blanket primary may have been able to maintain split delegations, more than it would produce them. That is, it was the re-election of long term incumbents from a minority party in a district that resulted in a bipartisan delegation.

    And most serious candidates had already decided to run before the Supreme Court decision had been handed down. Despite all this, there was a small increase in bi-partisan legislative delegations in 2008.

    And a focus on incumbents misses what happens when there is no incumbent, when there is an open seat. In such cases there tends to be a truly open race, open to all candidates and open to all voters.

  7. #4 Wasn’t Mike Foster a Democrat prior to 1995? And wasn’t he easily re-elected in 1999?

    He of course is the “Foster” in Foster v. Love. After the Supreme Court’s decision on the date of the open primary for congressional elections, the legislature considered various remedies. Foster and the House favored simply moving the primary to November (since non-federal elections in Louisiana are largely in odd-numbered years, they could keep the October primary for other elections). The Senate, and in particular Cleo Fields, favored switching back to the old 3-election partisan primary system.

    The legislature deadlocked, and then the federal district court which the case had been remanded to switched the date of the primary. This was challenged in the Love v. Foster where the plaintiffs argued that the remedy was to go back to the Louisiana law prior to institution of the open primary. The 5th Circuit upheld the district judge, and the Supreme Court did not take the case.

    Now that Louisiana has regressed to partisan primaries, candidates and voters are forced to go through up to three back-to-back elections, even to fill vacancies, and Louisiana has to use lockout devices on their voting machines to prevent voters who are voting in legislative elections from voting in congressional elections. The complexity of these lockout devices almost caused the Louisiana legislature to prevent 3rd parties from even participating in congressional elections.

  8. #7: In 1995, Mike Foster, age 65 and a lifelong Democrat, switched to the Republicans and was elected governor a few weeks later. Sorta like joining the church one Sunday and then getting elected chairman of the board of deacons the next Sunday. Yes, he was popular and had little opposition for re-election in 1999.

    State Sen. Cleo Fields, a former congressman and the second-place finisher for governor in ’95, spearheaded the wise restoration of party primaries for congressional elections.

    “Now that Louisiana has [restored] partisan [congressional] primaries, candidates and voters are forced to go through up to three back-to-back elections…”

    I forget what month the Bayou State’s regular congressional primaries are held in the new setup, but one could hardly say that the primaries, runoff primaries, and general election are “back-to-back.”

    Louisiana, of course, still uses the “top two” (“open primary”) for its state and local elections. I wish they would also restore party primaries for their state elections, but that doesn’t seem to be on the horizon.

    I believe that Washington state, if it wants to keep using the “top two” for its congressional elections, will wind up having to conduct the first round on the first Tuesday in November, with the second round at a later date. I also believe that Washington will be able to continue using some version of the “top two” for its state elections.

  9. #8 I think you need to go read Foster v Love again.

    Louisiana as a matter of law was declaring a winner of the open primary in October.

    And while it was not determinative as far as the legal issues, the opinion noted that Louisiana voters were forced to vote on two separate days for the final decision of President and Congress. One intent of the 1872 uniform date for congressional elections was to match the date set in 1845 for presidential elections.

    You can check the November 2008 results in Washington and find not only presidential and congressional elections, but gubernatorial and legislative elections as well.

    And it was certainly possible that voters could vote in Louisiana and then cross the river into Mississippi and vote again in November, this is not true in crossing the river between Washington and Oregon.

    In fact, the early partisan primaries in States like Illinois and Texas and Mississippi and late primaries like Hawaii would make it possible for a person to vote several times in decisive congressional elections (the one that functionally determines who sits in Congress).

    You don’t think the March primary in Mississippi introduces an unconstitutional residency requirement, and prevents the People from choosing their representative on the date set by Congress.

    And of course Congress has the authority to change the date of congressional elections. It did not exercise that authority for 83 years. It set a uniform date in 1872, but later relaxed it, and it wasn’t until 1960, 88 years later that all states were in compliance. Were the Supreme Court to make an adverse ruling against early primaries, Congress could change the law.

    Charlie Melancon in 2008 was no more elected in November than Billy Tauzin was in 1996. As a matter of law, if you wanted to vote for or against Rodney Alexander, you had to vote in October plus be a Republican.

Leave a Reply

Your email address will not be published. Required fields are marked *