In the wake of the Hackett-Brown fiasco, I keep reading comments claiming that contested primaries are a good thing and that the establishment is stupid for trying to prevent them. I first started hearing this theory after the Republicans swept most of the close Senate races in 2004, mainly due to turnout. The idea was that in states like Oklahoma and South Carolina, the winner had been able to raise his profile early due to the attention drawn by a contested primary, which emotionally invested the base in the eventual candidate and energized voters to turn out in droves come November.
The evidence for this theory seems pretty weak. Let's review some of the Senate races that year.
Let's consider two competing hypotheses:
Netroots hypothesis: Contested primaries are good because they raise the profile of anonymous candidates and energize the base to turn out in the general election.
Establishment hypothesis: Contested primaries are bad because they waste scarce resources on internecine battles and weaken the eventual nominee with mudslinging that depresses turnout in the general election.
To evaluate each hypothesis, let's use the presidential numbers for each state as a control. All else being equal, if a contested primary leads to a senatorial candidate outperforming the party's presidential candidate, I call that evidence that contested primaries are good. Conversely, if the winner of a contested primary subsequently underperforms the presidential candidate, I call that evidence that contested primaries are bad.
We'll look at all the open seats for 2004, plus Pennsylvania (because of Toomey) and minus Georgia (because that wasn't ever going to be close).
Colorado: (D) Contested (R) Contested
In Colorado, the establishment moderate Ken Salazar and the insurgent progressive Mike Miles contested the Democratic nomination. Miles pulled off some early wins, edging Salazar for the most delegates at the state Democratic convention, but Salazar roared back to crush Miles in the primary, 73-27. On the Republican side, Pete Coors and Bob Schaffer engaged in a brutal battle over who was more family-friendly (i.e. anti-gay), and the beer magnate Coors outvalued his opponent 61-39.
| | (D) | | | (R) |
President: |
Kerry | 47% |
| Bush | 52% |
Senate: |
Salazar | 51% |
| Coors | 47% |
Conclusion: kind of a wash. Salazar beat Coors in the general, 51-47, almost exactly reversing the margin by which Bush won Colorado. One might argue that Salazar's trial by fire against Miles toughened him up for the general. But one might also argue that Coors pyrrhic victory over Schaffer fatally weakened him in the general.
Florida: (D) Contested (R) Contested
In Florida, Betty Castor, Peter Deutsch, and Alex Penelas contested the Democratic primary. The primary went negative late in the campaign when Deutsch launched attack ads and Al Gore declared Penelas had been "treacherous and dishonest" during the 2000 recount. Castor won with 58% of the vote, though not before being tarnished with allegations that she had mishandled the case of Sami Al-Arian while she was president of the University of South Florida.
The Republican primary also turned nasty, pitting Mel Martinez against Bill McCollum. McCollum blasted Martinez for being a John Edwards trial lawyer, and Martinez responded by calling McCollum "the new darling of homosexual extremists". Martinez' broadside cost him the support of moderates and independents but won him Panhandle social conservatives. Martinez won the nomination, 42-33.
| | (D) | | | (R) |
President: |
Kerry | 47% |
| Bush | 52% |
Senate: |
Castor | 48% |
| Martinez | 50% |
Conclusion: Contested primaries are bad. Before the primary, Martinez was viewed as a respectable Hispanic moderate whose "reverse coattails" might push Bush over the top in this swing state. Instead, Martinez lagged Bush in the general election, an underperformance many attributed to his race to the right in the contested primary, which forced him to shed his moderate veneer.
Illinois: (D) Contested (R) Uncontested
In Illinois, the Democratic nomination was contested by Blair Hull, Dan Hynes, and Barack Obama. Obama won the nomination after frontrunner Hull spontaneously combusted over domestic abuse allegations. The Republican primary was uncontested, nominating Jack Ryan. After Ryan spontaneously combusted over domestic abuse allegations, the nomination actually became anti-contested, forcing the state party to recruit carpetbagger Alan Keyes, whose own spontaneous combustion turned out to be worse than having no candidate at all.
| | (D) | | | (R) |
President: |
Kerry | 55% |
| Bush | 44% |
Senate: |
Obama | 70% |
| Keyes | 27% |
No conclusion. While the contested primary undoubtedly honed Obama's skills for the uncontested general election, it's hard to draw any conclusions from a race in which one candidate can apparently cause his rivals to spontaneously combust. Update 2:Conclusion: Contested primaries are good. It's been pointed out to me that the contested primary selected an outstanding candidate and eliminated an unworthy frontrunner, while an uncontested process was responsible for two disastrous nominees in a row.
Louisiana (D) Contested (R) Uncontested
Louisiana has a unique open primary/general election in which multiple candidates from both parties compete in the same general election. An outright majority wins the election; otherwise, the two candidates with the most votes meet again in a December runoff. The Democrats put up three candidates: Chris John, John Kennedy, and Arthur Morrell. Republicans united behind David Vitter.
| | (D) | | | (R) |
President: |
Kerry | 42% |
| Bush | 57% |
Senate: |
John | 29% |
| Vitter | 51% |
| Kennedy | 15% |
| Morrell | 2% |
Conclusion: Contested primaries are bad. Vitter managed an outright win, narrowly avoiding a runoff. One wonders if he could have done this against a single, undistracted Democratic rival, especially given Louisiana's affinity for ticketsplitting.
Oklahoma: (D) Uncontested (R) Contested
The Democratic primary was not seriously contested since Brad Carson was a dream candidate. In the Republican primary , lightly funded insurgent rightwing darling Tom Coburn upset establishment candidate Kirk Humphreys 53-25. Since Coburn defines the phrase "batshit crazy", his victory instantly converted deep red Oklahoma into a first-tier Senate race.
| | (D) | | | (R) |
President: |
Kerry | 34% |
| Bush | 66% |
Senate: |
Carson | 41% |
| Coburn | 53% |
Conclusion: Contested primaries are bad. Despite (or perhaps because of) widely publicised verbal gaffes (e.g. abortion doctors should be charged with capital crimes, high school locker rooms are filled with lesbians), Coburn beat Carson by 12 points, which sounds good till you hear that Bush beat Kerry by a whopping 32 points. On the other hand, a win is a win, and by challenging the primary, the Republican grassroots did manage to seat their dream candidate in the Senate.
Pennsylvania: (D) Uncontested (R) Contested
In Pennsylvania, the rabidly right wing Club For Growth pitted Pat Toomey against RINO incumbent Arlen Specter in the Republican primary. Specter barely eked out a win over Toomey, 51-49, and was noticeably weakened going into the general election against Democrat Joe Hoeffel.
| | (D) | | | (R) |
President: |
Kerry | 51% |
| Bush | 49% |
Senate: |
Hoeffel | 42% |
| Specter | 53% |
Conclusion: Contested primaries are bad. Specter pulled out a convincing win, mainly due to crossover from moderate Democrats who had been supporting him for decades. The race was closer than it could have been because he was forced to make concessions to the right wing that cost him a lot of independent cred. On the other hand, a lot of swing voters apparently figured that if the right wing hated him that much, he couldn't be that bad. Update: As the only incumbent on my list, I should probably throw up Specter's previous reelection numbers. 1998: Specter 67%, Lloyd 31%; 1992: Specter 49%, Yeakel 46%, Perry 5%; 1986: Specter 56%, Edgar 43%. (1992 was a tight squeeze for Specter: he lost support on the left due to Anita Hill and he lost support on the right to an anti-choice Libertarian. Presidential numbers that year were Clinton 45%, Bush 36%, Perot 18%.)
North Carolina: (D) Uncontested (R) Uncontested
In North Carolina, Republican nominee Richard Burr and Democratic nominee Erskine Bowles both won uncontested primaries.
| | (D) | | | (R) |
President: |
Kerry | 44% |
| Bush | 56% |
Senate: |
Bowles | 47% |
| Burr | 52% |
No conclusion. Burr rode Bush's coattails to victory in the general.
South Carolina: (D) Uncontested (R) Contested
In South Carolina, Inez Tenenbaum won an uncontested Democratic nomination. The Republican primary was a three-man, two-round contest in which Jim DeMint emerged victorious and not particularly damaged.
| | (D) | | | (R) |
President: |
Kerry | 41% |
| Bush | 58% |
Senate: |
Tenenbaum | 44% |
| DeMint | 54% |
Conclusion: Contested primaries are good. Although DeMint underperformed Bush in his general election victory, the difference was probably due to his foot-in-mouth endorsement of a national sales tax late in the campaign. The race was probably only as close as it was because Tenenbaum was able to campaign as a moderate in this red state.
Conclusion
In conclusion, contested primaries unambiguously hurt candidates in Florida, Louisiana, and Pennsylvania. Update 2: Contested primaries were unambiguously good in South Carolina and, especially, Illinois.
In Oklahoma, the grassroots got everything they wanted by challenging the primary, and got away with it at the cost of making the race much closer than it could have been. In Pennsylvania, the grassroots got half of what they wanted, despite Toomey's loss, by intimidating Specter into better behavior in the new Congressional session; nevertheless, it was an irresponsible risk, as Toomey would almost certainly have been defeated in the general, unnecessarily surrendering a Republican seat in a purple state.
So there's a reason conventional wisdom says that contested primaries are bad. Having said that, I will now claim that every rule has exceptions and that the exception to this rule is named Ned Lamont.
Update 3: I changed the title to clarify that I'm not claiming that contested primaries are unambiguously bad, but rather that they aren't unambiguously good. A contested primary is an extra opportunity for more people to drive up the negatives of the eventual nominee. That's an affordable luxury in deep red states like Oklahoma or deep blue states like Connecticut. In closer races, it's not as easy a call.