When I awoke on Wednesday, 3 November 2010, having been unable to stay awake to await the late returns from MN-08, I was extremely frustrated to learn of the defeat of United States Representative Jim Oberstar, the Democratic-Farmer-Labor (DFL) incumbent, to Republican challenger Raymond "Chip" Cravaack. Oberstar's record of service to the MN-08 is both long and distinguished, and while the polling data in the few weeks prior to the 2010 midterm election indicated that this race was going to be close, I must nevertheless admit that Cravaack's victory came as a both a surprise and a profound disappointment. This district, after all, has been held by Oberstar since 1975, has been in DFL hands since 1947 and is described in dKosopedia as "one of the safest Democratic seats in Congress" (last edit: 13 December 2006).
In light of a discussion I had with a fellow Duluthian Kossack in a comment-thread on the day after the election, in which we noted anecdotally some of the geographic aspects of Oberstar's and Cravaack's support as indicated by the presence and absence of candidates' lawn-signs, I decided to explore the election results in more depth. I will use two diaries to present my observations on the geography--the "anatomy"-- of Oberstar's defeat.
INTRODUCTION TO MN-08
For those unfamiliar with Minnesota's Eighth Congressional District, it encompasses seventeen counties in the northeast portion of the state (Aitkin, Carlton, Cass, Chisago, Cook, Crow Wing, Hubbard, Isanti, Itasca, Kanabec, Koochiching, Lake, Mille Lacs, Morrison, Pine, St. Louis and Wadena) as well as the southeast quadrant of Beltrami County. Essentially, this district covers Minnesota’s Iron Range and Arrowhead landforms. To the north is Canada, to the east Lake Superior and Wisconsin, to the west MN-08 abuts MN-07, and to the south it abuts MN-06, home of the crazy-eyed Michele Bachmann. The most populous urban center in MN-08 by far is Duluth, in southeastern St. Louis County, with a population of circa 85,000. The second tier of urban centers includes Hibbing (circa 17,000), Brainerd (circa 13,000), Cloquet (circa 11,000), Virginia (circa 9,000), Grand Rapids (circa 8,000), Little Falls (circa 8,000) and International Falls (circa 7,000). Voting data for cities and towns will be discussed in the context of precinct-level data.
MAPPING THE 2010 VOTE IN MN-08
My presentation and analyses of the 2010 election data for MN-08, are at three geographic levels: district-wide results, county-level results and precinct-level results. This discussion is far, far too long for a single diary, so I will use this first diary (Part I) to address the district-wide and county-level results, then post a second diary (Part II) early next week to address the precinct-level results. All voting data, both for 2010 and for previous elections, have been obtained from the website of the Office of the Minnesota Secretary of State. For presentation and analysis, these voting data have been joined within Geographic Information Systems (GIS) software to geographic boundary files available from the Geographic Information Services division of Minnesota’s Legislative Coordinating Commission. All of these data are thus in the public domain, and available free of charge.
DISTRICT-WIDE RESULTS, 2010
There were four candidates on the ballot for United States Representative in Minnesota's Eighth Congressional District: DFL incumbent Jim Oberstar, Republican Chip Cravaack, Independence Party candidate Timothy Olson and Constitution Party candidate Richard (George) Burton. District-wide results reported by the Office of the Minnesota Secretary of State for each candidate's total votes and percentage of the total ballots cast are listed below. Also included is a statistic estimating "voter intensity" in MN-08. This value is calculated by dividing the total number of votes cast by the total number of registered voters (pre-registered and election-day registration). The "voter intensity" measure should not be mistaken for a direct measure of voter turnout, for which we would need to divide the total number of votes cast by an official estimate of the voting-age population at all geographic levels (state, district, county and precinct). The official Census estimate of the voting-age population is only available at the geographic level of the state, thus accurate measures of voter turnout are only reliable at the level of the state. The measure of "voter intensity," while not linked directly to voter turnout as a percentage of the total voting=age population, nevertheless provides a proxy statistic enabling the comparison of voting rates at more detailed levels of geography than the state.
2010 Midterm Election (voter intensity = 66.35%; MN turnout = 55.46%)
Candidate | Votes | Percentage |
---|
Chip Cravaack | 133,486 | 48.18% |
Jim Oberstar | 129,084 | 46.59% |
Timothy Olson | 11,876 | 4.29% |
Richard Burton | 2,492 | 0.90% |
Write-In | 132 | 0.05% |
When we survey articles and blog-postings offering commentary on Oberstar’s defeat, we encounter the following explanations most frequently. First, the general "anti-incumbent mood," the "wave," of the 2010 election season nationally was in effect. With respect to Oberstar specifically, there is a perception among the pundits and pols that the voters of MN-08 found Oberstar to have become too much of a "Washington insider" during his tenure in the House, was too willing to back expensive federal projects, particularly through his chairmanship of the House Committee on Transportation and Infrastructure, and was too much a champion of earmarks. In short, the perception of voter intent in MN-08 is that a younger, fresh face with a smaller-government perspective was needed to represent the district.
Second, there is a perception that the national "enthusiasm gap" was in play in the district, the effects of which were magnified in a midterm election year when voter intensity (yes, even in stalwart turnout-leading Minnesota...) is on the order of ten to fifteen percentage-points lower than in presidential election years. Oberstar himself noted the lower-than-expected rate of voter turnout in traditionally Democratic counties, especially St. Louis County, in his concession speech.
Third, there has been a widening partisan divide within the district, manifest in two dimensions: both along a north / south axis, with the south trending increasingly conservative (Cravaack is himself from Lindstrom in Chisago County in the far southeast of MN-08), as well as among rural and urban precincts, with the former trending increasingly conservative. My intent over the course of this two-diary series is to shed some light on these potential factors in Oberstar’s loss, and introduce several other aspects of the race that seem, to my eyes, particularly significant.
DISTRICT-WIDE RESULTS, 2002–2010
As noted in the Introduction to this diary, MN-08 has been held by the DFL since 1947. Representative John Blatnik held the seat from 1947 to 1974, and Representative Jim Oberstar has held the seat since then. Based on voting data from the 2004 and 2008 presidential elections in comparison with national results, MN-08 has been assigned a Cook Partisan Voting Index (CPVI) of D+3. This is certainly not the most partisan of Congressional Districts in Minnesota, but over the past few election cycles MN-08 has had a definite Democratic lean. Below are the district-wide results (total votes and percentage of votes) from the 2002 to 2010 elections, including both a measure of "voter intensity" for the district and the official measure of voter turnout in Minnesota from the Secretary of State.
2002 Midterm Election (voter intensity = 72.47%; MN turnout = 64.89%)
Candidate | Votes | Percentage |
---|
Oberstar (DFL) | 194,909 | 68.65% |
Lemen (GOP) | 88,673 | 31.23% |
Write-In | 349 | 0.12% |
2004 Presidential Election (voter intensity = 80.05%; MN turnout = 78.77%)
Candidate | Votes | Percentage |
---|
Oberstar (DFL) | 228,586 | 65.22% |
Groettum (GOP) | 112,693 | 32.15% |
Presley (Green) | 8,933 | 2.55% |
Write-In | 271 | 0.08% |
2006 Midterm Election (voter intensity = 66.05%; MN turnout = 60.46%)
Candidate | Votes | Percentage |
---|
Oberstar (DFL) | 180,670 | 63.61% |
Grams (GOP) | 97,683 | 34.39% |
Welty (Unity) | 5,508 | 1.94% |
Write-In | 155 | 0.05% |
2008 Presidential Election (voter intensity = 78.63%; MN turnout = 78.49%)
Candidate | Votes | Percentage |
---|
Oberstar (DFL) | 241,831 | 67.69% |
Cummings (GOP) | 114,871 | 32.15% |
Write-In | 582 | 0.16% |
2010 Midterm Election (voter intensity = 66.35%; MN turnout = 55.46%)
Candidate | Votes | Percentage |
---|
Cravaack (GOP) | 133,486 | 48.18% |
Oberstar (DFL) | 129,084 | 46.59% |
Olson (IND) | 11,876 | 4.29% |
Burton (CON) | 2,492 | 0.90% |
Write-In | 132 | 0.05% |
In graphic form, here is a chart of the percentages of the district-wide vote won by Oberstar and his Republican challenger for the 2002 to 2010 elections.
The differences in percentages in these elections are as follows:
Year | Margin |
---|
2002 | Oberstar +37.42% |
2004 | Oberstar +33.07% |
2006 | Oberstar +29.22% |
2008 | Oberstar +35.54% |
2010 | Oberstar -1.59% |
While the sample of elections is admittedly too small for any kind of serious statistical analysis, it is interesting nonetheless that the percentage-differences of the district-wide vote won by Oberstar from 2002 to 2010 correlates moderately to the measure of voter intensity.
Year | DFL Margin | Voter Intensity |
---|
2002 | +37.42% | 72.47% |
2004 | +33.07% | 80.05% |
2006 | +29.22% | 66.05% |
2008 | +35.54% | 78.63% |
2010 | -1.59% | 66.35% |
In other words, and recognizing that it would not be prudent to press this point too hard given such a small sample, relatively high turnout among registered voters in MN-08 does seem to correlate to some degree with DFL success in the district. Since Minnesota does not include a declaration of political-party affiliation as part of voter registration, we must rely almost solely on exit-polling for information regarding voter intensity, and voter turnout, by party. We will see below whether the relationship between voter intensity and DFL success holds at the level of the county and the precinct, for which sample-size is less of an issue.
COUNTY-LEVEL RESULTS, 2010
At a somewhat more detailed level of geographic analysis, we can assay the results of voting data in MN-08 at the level of the county. Here, in tabular form, are the percentages of the 2010 vote won by Oberstar and by Cravaack in the eighteen counties of MN-08:
County | Oberstar | Cravaack |
---|
Aitkin | 43.64% | 50.04% |
Beltrami | 42.36% | 51.90% |
Carlton | 53.79% | 42.07% |
Cass | 38.65% | 54.25% |
Chisago | 38.26% | 55.59% |
Cook | 55.20% | 40.05% |
Crow Wing | 38.53% | 54.68% |
Hubbard | 38.72% | 55.02% |
Isanti | 36.80% | 56.63% |
Itasca | 48.24% | 46.35% |
Kanabec | 39.26% | 53.80% |
Koochiching | 52.54% | 42.74% |
Lake | 54.92% | 41.81% |
Mille Lacs | 40.05% | 52.90% |
Morrison | 35.79% | 57.82% |
Pine | 43.37% | 51.02% |
St. Louis | 56.92% | 39.66% |
Wadena | 37.97% | 55.97% |
In a very general way, we can note that Oberstar won six counties in the district, while Cravaack won twelve. We can also note the average percentages associated with counties won by and lost by Oberstar (avg. win = 53.60%; avg. loss = 39.45%), and those won by and lost by Cravaack (avg. win = 54.14%; avg. loss = 42.11%). These data reflect that Cravaack tended to win by a greater margin, while Oberstar tended to lose by a greater margin. This detail may speak both to the perception of an "anti-incumbent mood" in MN-08 as well as to the increasingly bloc-like partisan geography of the district.
More illuminating with respect to the formation of partisan blocs in the district is the following visualization of these county-level results, with counties won by Oberstar represented in blue, those won by Cravaack represented in red and the margin-of-victory in percentage-points indicated below the name of the county.
The regional distribution of county-level support for Oberstar versus support for Cravaack is immediately apparent. Counties won by Oberstar form a contiguous area in the northeast and north, representing Duluth, the Iron Range and the Arrowhead. Cravaack's support is represented by a contiguous area of counties in the south (adjacent to MN-06, Tea Party favorite Michele Bachmann's district) and the west (adjacent to MN-07, Blue Dog Collin Peterson's district).
This partisan geography within MN-08 is not restricted to support for either Oberstar or Cravaack in the election for United States Representative, but is also apparent in the composition of the Minnesota State Senate and House in the wake of the 2010 election, as the following image demonstrate.
The county-level data for MN-08 do seem to confirm that there has been a widening partisan divide within the district, broadly in line with the partisan representation of Minnesota’s State Senate and House. There is a clearly pronounced DFL stronghold in the northeast, while the south and west are trending conservative, perhaps through affinities to MN-06 in the south and MN-07 in the west. Indeed, given that the average margin of victory for Oberstar in the urban precincts of Duluth itself was a remarkable 16.14%, we might intuit, as several local pundits have in the wake of the election, that support for Oberstar in 2010 diminished as a function of distance from Duluth. That statement, however, as intuitively "correct" as it may feels to liberal Northlanders such as myself, does not necessarily hold true when we consider the measures for "voter intensity" as the level of the county. The following image shows counties won by Oberstar in blue, those won by Cravaack in red and the measure of voter intensity—the ratio of total votes cast to the total number of registered voters—indicated below the name of the county.
Note that St. Louis County, which includes Duluth, has the absolute lowest measure for voter intensity, at 62.56%. For those unfamiliar with the population distribution within MN-08, this may seem a fairly insignificant detail. However, if we observe that St. Louis County contains approximately 35% of the total population of MN-08, and that Crow Wing County, the second largest by population, contains just 8%, then I believe that the significance of relatively low voter intensity in St. Louis county, a traditional DFL stronghold, is more readily apparent. In effect, a drop of 1% for voter intensity in St. Louis County is equivalent to drops of far larger percentages elsewhere in the district. I will expound upon the ramifications of this observation in a discussion of voter intensity by precinct in the second diary of this series.
COUNTY-LEVEL RESULTS, 2002–2010
From 2002 to 2008, Jim Oberstar won all eighteen counties in MN-08 in every election. The following table provides Oberstar’s margin-of-victory in percentage-points over his Republican challenger in these races, as well as the 2010 margins.
County | 2002 | 2004 | 2006 | 2008 | 2010 |
---|
Aitkin | 29.11% | 24.57% | 21.97% | 27.68% | -6.40% |
Beltrami | 6.35% | 9.70% | 4.58% | 19.29% | -9.54% |
Carlton | 52.72% | 50.23% | 44.73% | 50.55% | 11.72% |
Cass | 15.17% | 16.11% | 10.28% | 18.85% | -15.60% |
Chisago | 15.12% | 14.94% | 18.14% | 18.92% | -17.33% |
Cook | 40.38% | 37.51% | 42.30% | 44.73% | 15.15% |
Crow Wing | 20.59% | 15.79% | 11.34% | 21.05% | -16.15% |
Hubbard | 10.39% | 9.99% | 7.45% | 14.69% | -16.30% |
Isanti | 15.64% | 14.80% | 10.42% | 17.36% | -19.83% |
Itasca | 39.45% | 35.14% | 31.82% | 36.01% | 1.89% |
Kanabec | 24.33% | 22.95% | 16.39% | 20.71% | -14.54% |
Koochiching | 50.24% | 35.43% | 33.65% | 45.13% | 9.80% |
Lake | 54.74% | 48.99% | 46.58% | 50.71% | 13.11% |
Mille Lacs | 21.54% | 19.60% | 16.72% | 22.69% | -12.85% |
Morrison | 28.36% | 26.03% | 8.54% | 23.36% | -22.03% |
Pine | 35.07% | 34.05% | 29.77% | 33.33% | -7.65% |
St. Louis | 59.57% | 51.58% | 49.86% | 54.51% | 17.26% |
Wadena | 8.88% | 3.88% | 0.43% | 10.41% | -18.00% |
In graphic form, here is a chart of Oberstar’s margins at the level of the county for the 2002 to 2010 elections.
These data lend credence to the 2006 dKosopedia description of MN-08 as "one of the safest Democratic seats in Congress," an opinion that I must admit was shared by myself and many of my Democratic acquaintances in the district until just a few weeks prior to the 2010 election, and now lies in tatters.
Let us return, for a moment, to the observation above regarding voter intensity in the 2010 election, and contextualize those values by charting them in relation to the measures for voter intensity by county in the 2002 to 2008 elections.
County | 2002 | 2004 | 2006 | 2008 | 2010 |
---|
Aitkin | 75.21% | 82.12% | 72.55% | 80.70% | 69.71% |
Beltrami | 69.50% | 77.80% | 69.21% | 79.64% | 70.47% |
Carlton | 77.28% | 85.48% | 66.90% | 79.02% | 65.95% |
Cass | 71.98% | 79.79% | 65.59% | 78.45% | 69.27% |
Chisago | 72.73% | 78.35% | 64.16% | 79.10% | 68.56% |
Cook | 82.67% | 88.27% | 79.85% | 85.83% | 78.88% |
Crow Wing | 75.06% | 80.88% | 69.09% | 79.98% | 68.03% |
Hubbard | 76.70% | 81.81% | 72.05% | 79.26% | 72.18% |
Isanti | 71.43% | 82.40% | 64.42% | 79.61% | 64.52% |
Itasca | 73.06% | 81.55% | 69.56% | 81.19% | 70.69% |
Kanabec | 71.81% | 80.40% | 66.97% | 79.81% | 67.85% |
Koochiching | 70.04% | 79.03% | 67.94% | 79.45% | 65.91% |
Lake | 75.75% | 82.70% | 72.45% | 80.77% | 71.26% |
Mille Lacs | 71.75% | 79.75% | 65.46% | 78.64% | 65.91% |
Morrison | 74.42% | 78.00% | 67.69% | 78.66% | 67.33% |
Pine | 69.29% | 78.19% | 64.42% | 73.99% | 65.88% |
St. Louis | 70.62% | 78.94% | 62.88% | 77.21% | 62.56% |
Wadena | 73.14% | 78.53% | 68.92% | 80.53% | 66.61% |
Above, and echoing Oberstar’s comments during his concession speech, I proposed the somewhat counter-intuitive observation that relatively low voter intensity in St. Louis County may have been a significant factor in the loss of MN-08, in spite of the fact that the county was won by Oberstar with a margin-of-victory of 17.26% over Cravaack. In recent historical context, the measure of voter intensity for St. Louis County has been consistently below the average measure for counties within the district.
Year | Average | St. Louis | Difference |
---|
2002 | 73.47% | 70.62% | -2.85% |
2004 | 80.78% | 78.94% | -1.84% |
2006 | 68.34% | 62.88% | -5.46% |
2008 | 79.55% | 77.21% | -2.34% |
2010 | 68.42% | 62.56% | -5.86% |
The difference in the measure of voter intensity for St. Louis County, although below average in comparison to all of the eighteen counties within MN-08 for all of these election cycles, is most exaggerated in midterm election years (2002, 2006, 2010) and deviates most strongly from the average in 2010. Without "giving away the game," I will foreshadow later analyses of the precinct-level data within St. Louis County by noting that of the fourteen precincts in the county with a measure of voter intensity less than 50.00%, twelve are urban precincts in Duluth. As a Duluthian myself, and one who has in the past lauded somewhat pridefully the political engagement of my fellow Duluthians, these statistics are profoundly disturbing.
SUMMARY: DISTRICT-WIDE AND COUNTY-LEVEL RESULTS
The "red meat" of the data-analysis for the 2010 election of Tea Partier Lite Raymond "Chip" Cravaack in MN-08, and the surprising defeat of incumbent DFLer Jim Oberstar, of course lies in the precinct-level data. Even so, cursory analyses of the district-wide and county-level data do document two of the factors in play.
First, the partisan regional geography of MN-08 seems to be hardening, with the south and west trending increasingly conservative. The southern counties—in particular Chisago, Isanti, Pine as well as southern Mille Lacs—are essentially exurbs of the Twin Cities, and have far more in common demographically with Michele Bachmann’s MN-06 than with Duluth, the traditional core of MN-08. The western counties, in my opinion, seem to be sympathetic to the Blue Dog vision of the DFL as represented by Collin Peterson of MN-07, who easily retained his seat in 2010.
Second, voter turnout, proxied here by a measure of voter intensity, without doubt worked to Cravaack’s advantage. Relatively high voter intensity in the western counties (Beltrami, Hubbard and Cass) and in Cravaack’s home county of Chisago in the south, coupled with relatively low voter intensity in St. Louis County, the most populous county in MN-08, seem to have swung the pendulum in Cravaack’s favor. It is worth recalling that early on Election Night, awaiting anticipated (and precedented) positive results from precincts in St. Louis County, Oberstar gave a kind of victory speech, declaiming thus:
I want the nation to know that here in the northland of Minnesota, that [anti-incumbent, Tea Party] wave has crashed on the rocks of Minnesota decency, and goodness, and caring, and justice for all citizens!
By 3:00 AM, when the results from late-reporting precincts in St. Louis County were reported, the tide had changed.
In Part II of this diary-series, which I will post early next week, I will present and discuss the precinct-level data for the 2010 race for United States Representative in MN-08. At the level of the precinct we will be able to address the hypothesis of an increasingly partisan divide among rural and urban precincts, with the former perceived to be trending increasingly conservative. We will also be able to model in more detail the relatively low voter intensity in St. Louis county, as well as the relatively high voter intensity in the south of MN-08, particularly in the conservative exurbs of the Twin Cities.
In some ways, I suppose I should apologize for the length of this diary and apologize preemptively for the length of the second installment. At the same time, however, I feel that what may be taken for a long-winded exercise is both cathartic (personally) and instructive. I am trying very hard to come to terms with the defeat of Jim Oberstar, my Congressman, and I am attempting to identify some of the challenges that we Northland DFLers must recognize, and confront, for 2012 and beyond. I ask, then, your indulgence.