Primary reconfirms Obama's rural MO weakness

As I scoured the numbers from last week's "meaningless" Missouri presidential primary, I was startled to find that the real news was not in the hyped Republican contest but in the (technically meaningful) Democratic contest.

Overall, President Obama received a slightly underwhelming 88.3% of the vote in the Democratic primary against three little known vanity candidates and an option to send uncommitted delegates to the convention. Uncommitted, at 6.3%, was the most popular Democratic alternative to Obama. In contrast, incumbent President George W. Bush received 95.1% of his party's vote in an analogous Missouri primary in 2004. No Democrat wants to underperform Dubya!

However, it's the geography behind the numbers that should provide greater concern for the President and his strategists.

Obama was predictably strong in urban areas (the cities of St. Louis and Kansas City plus St. Louis County) where African American voters (Obama's strongest demographic) dominate the vote in the Democratic primary, plus Boone County, where faculty and some students at the state's largest college campus predominate. Obama polled a near-unanimous 96% in St. Louis and Kansas City, 93.3% in St. Louis County, and 92.8% in Boone County.

Unlike St.Louis County, Democrats in Kansas City's suburbs showed less love for Obama, with Jackson County (excluding Kansas City itself), Clay County to the north and Cass County to the south giving the President a little less than his share of the vote in the state as a whole.

But the President's numbers were more concerning in rest of the state. In rural and exurban areas (everything but St. Louis City and County, Kansas City, the rest of Jackson County and Boone County), 17.2% of Democratic Primary voters voted against their party's president. The non-Obama Democratic vote topped 20% in nearly half the rural counties, and over 30% in seven of them. In Reynolds County in the southern Missouri lead belt, it came within one vote of 40%. Even in relatively populous Buchanan County (St. Joseph and environs), the non-Obama Democratic vote was 24.3%.

Democrats have performed poorly outstate in recent November general elections, but the above numbers are from a Democratic Primary! Those numbers aren't significantly tainted by crossover votes, because nearly all of the Republican and independent voters who drive the general election numbers were most likely drawn to the highly contested and well publicized Republican contest, if they voted at all. The voting pool here would have been almost entirely true Democrats, and over a sixth of them in the rural and exurban half of the state said no to their President.

Not everyone who votes against his party's President in a primary votes for the other party's candidate in the following general election. But some (including many who chose not to vote at all in the primary) may exercise their frustration by not voting at all in November. That could adversely affect Sen. Claire McCaskill (who shares much of the President's record) and the rest of the Democratic ticket.

Republican Primary

The contest that I expected to analyze turned out to be pretty homogeneous. There was no significant urban/rural split, as former Sen. Rick Santorum carried every county, including the separately tabulated cities of St. Louis and Kansas City, defeating establishment-endorsed runner-up Mitt Romney by 30 points statewide. Santorum's success here was aided by the absence of conservative rival Newt Gingrich on the ballot, but Santorum also won caucuses the same day in two other states where Gingrich did compete.

Here are what few Republican patterns I could discern:

• Anti-war libertarian Ron Paul (12.2% statewide) ran strongest in the big cities, winning 20.3% in the City of St. Louis (compared to just 10.6% in St. Louis County) and 15.2% in Kansas City, although his best county was Mercer in northern Missouri, where Paul's 27.9% was good for second place over Romney. Paul had been expected to overperform in academic centers, but only Boone County (University of Missouri) met expectations with 18.2%. Paul ran marginally ahead of his statewide share in Phelps (Missouri S&T), Adair (Truman State) and Nodaway (Northwest Missouri State) Counties, but trailed in Johnson County (University of Central Missouri). Other campuses (e.g. Missouri State) are situated in counties in which their share of the vote is insignificant. Paul's relative strength appeared to come at Santorum's expense rather than Romney's.
  
• Romney performed relatively well (30%+) in the Kansas City and St. Joseph areas in western Missouri. Santorum's victory margin over Romney in Buchanan County was single digits. Romney also flirted with 30% (29.8) in St. Louis County, where he enjoyed fundraising success.
• Santorum's share (55.2% statewide) was quite consistent across the state. The biggest departure from the norm was Ralls County, south of Hannibal, which delivered 75.4% for Santorum. He topped 60% in some exurban counties (Franklin west of St. Louis, and Cass and Lafayette south and east of Kansas City), but not in others. He marginally underperformed his statewide share in urban areas, primarily due to Paul's relative strength there.