Two economics professors at Gonzaga University in Spokane have published “The Effect of the Top Two Primary on the Number of Primary Candidates.” It will appear in a future issue of the Social Science Quarterly, journal of the Southwestern Social Science Association. It is already available on-line, but it requires payment. See here for the link to the article on the Wiley Online Library.
The article is by John H. Beck and Kevin E. Henrickson. Comparing the Washington state classic open primaries of 2004 and 2006 with the top-two primaries of 2008 and 2010, the article conclude that the top-two system appears to have caused a reduction in the number of Democrats who run for the Washington state legislature. Each year, there are 123 or 124 regularly-scheduled legislative races in Washington state, and the study uses complex statistical analysis to show that, in the average election year under top-two, 18 to 19 fewer Democrats run for the legislature than if the top-two system did not exist. The evidence in the article is entirely statistical, except for one anecdote, in which the chair of the Democratic Party is quoted as saying, “I, as party chair, have to go and talk people into not participating, and I think that’s really unfortunate.”
The reason party leaders discourage party members from running is that if a major party has too many candidates for a single seat, the party is in some danger that no member will qualify for the November ballot. The article thus provides evidence for the point that top-two open primary systems reduce voter choice in primaries, and enhance the power of “party bosses”. Thanks to Mark Rogalski for news of the article.
Gee – top 2 primary machinations akin to the EVIL robot party hack boss stuff before 1888 (the official primary) ???
—-
P.R. and nonpartisan App.V.
NO primaries
Comparing 2006 with 2010, didn’t the number of Democratic candidates increase from 120 to 131?
There are typically more candidates in a race with no incumbent, perhaps the study is somehow trying to take that into account. That is, is the study is using complex statistical legerdemain to determine couldas and wouldas, and failed to note the didas that can be determined by simple subtraction? 131 – 120 = net gain of 11 from 2006 to 2010.
I wonder what assumptions are being made. Is the paper available for less than $35?
The number of Democratic House candidates was/is:
2004: 118
2006: 93, -25
2008: 101, +8
2010: 106, +5
2012: 108, +2
I would be wary of including senate races since they are for 4-year terms, the areas covered by the two sets of senate districts are quite different geographically and politically, making 2-year trends useless. In addition, some mid-term special elections might be included, which because there is no incumbent, would likely increase the number of candidates seeking election. This could mask trends, or be perceived as a trend, when the vacancy was an anomalous event.
The legislature passed Top 2 in early 2004, while the blanket primary was still being litigated. It also included a backup for the Pick-A-Party primary. After Gary Locke’s misguided line item veto of the Top 2 primary, the Pick-A-Party primary was imposed for 2004 (when 10% of actual voters failed to register a preference in the gubernatorial primary). Serious candidates for the legislature typically do not wait until late spring to decide to run. Therefore, it is dubious to assume that the number of Democratic candidates in 2004 was an effect of the Pick-A-Party primary.
The low point in the number of Democratic candidates for the House was in 2006, following 2 years of imposition of the Pick-a-Party primary.
Similarly, the US Supreme Court decision was not until March 2008, and the Forces of FUD attempted to prevent its implementation on remand to the lower courts. So while the number of candidates did increase, most candidates may have anticipated that 2008 would have been yet another Pick-a-Party primary.
The low point was in 2006 during the second use of the Pick-a-Party primary, and Democratic candidacies have climbed since then.
Relatively few primaries have had 3 or more Democratic candidates. Since 2004, there have been 15 House races among the 490 total that have had 3 or more races (3.1%). Yet these rare races represent 61 of the 524 candidacies (11.6%). Typically, a large number of candidates are attracted to an open seat in a district dominated by their party. Under the old system, if they could finish first among Democrats in the primary they were assured of election. The same is almost true now. If they finish first among Democrats, they are assured a place in the Top 2. If their opponent is a Republican they will win the general election. If their opponent is another Democrat, they simply have to prove that they are the favored Democrat from among the entire electorate. The only candidates discouraged from running in such a race are Democrats who are unable or unwilling to engage Republican or independent voters.
The number of races with a large number of Democrat candidates does not depend on whether a blanket, pick-a-party, or Top 2 primary is used. It depends on an open seat in a Democratic-dominated area.
An example is LD 36 in Seattle. Mary Lou Dickerson was elected to Representative Position 2 in 1994, winning a 6-candidate primary when the former representative switched to the senate, where continues to serve. After a weak primary challenge in 1996, Dickerson has not been challenged in the primary and has rolled up 80% margins when the Republicans have bothered to put up a candidate.
After 9 elections and 18 years in the legislature Dickerson is retiring, and 5 Democrats, 1 Republican, and 1 Progressive have filed. The Democrats did not file because it was a blanket, Pick-A-Party, or Top 2 primary. They filed because it is an overwhelmingly Democratic district with no incumbent.
The number of races with 2 or 3 or more Democratic candidates has increased since 2004:
2004: 4 (3+), 14 (2)
2006: 2, 2
2008: 1, 8
2010: 5, 12
2012: 4, 15
In other words, the number of races where the menu has included a choice among Democrats has increased.
The number of races where there have been no Democrats on the primary ballot has also increased, while the number of races with a single Democratic candidate has decreased.
2004: 4 (0 candidates), 77 (1 candidates)
2006: 14, 80
2008: 7, 82
2010: 19, 62
2012: 17, 62
That is, the reason for slightly fewer Democratic candidates in 2012 compared to 2004 is not due to multiple candidates not filing for fear of splitting the party vote. In fact such candidacies have increased. But rather it is in districts, mostly east of the Cascades, where no Democratic candidates have filed.
In 2012, there are 5 districts where no Democrat filed for either House seat, including 3 districts where there is no Senate candidate. The reason for
There are also 7 districts where there is a Democratic candidate for one position, but not the other. In some cases, this may be tactical. There are some deliberate split-ticket voters. Under Washington’s system of two House races per legislative district, it may be advantageous to have a candidate in only one race, since ticket-splitters will be forced to all support the same candidate.
If Dwight Pelz has attempted to discourage multiple Democratic candidacies, he has failed, since they have increased under Top 2. He may also have failed to encourage candidacies in areas where Democrats are weak, particularly east of the Cascades, or alternatively, his election as party chair is reflective of where Democrats are strongest. Most likely, his comments are self-serving to justify continued litigation by his party.