Bernie Sanders is widely expected to win the Oregon primary coming up in mid-May. Hillary Clinton lost the state by double digits in 2008, and though there have been few polls, they have tended to show Sanders doing well here, if not winning. It is a very heavily white state, like most of the states he has won so far, and the electorate is even whiter than the general population. It is also a state filled with progressive voters, and where the Portland metropolitan area, which provides about half of the votes in statewide Democratic primaries, is dominated by counties which have some of the largest shares of voting-eligible young people in the state, and he has done very, very well with young voters. The next biggest metropolitan area in the state is the Eugene-Springfield area, which is dominated by a college town and is even whiter than the state as a whole. On paper it appears like it would be a very good state for Sanders.
While it is likely that he will win the primary, I think some are expecting a much bigger win than is likely. Sanders’ 72% in Washington and 79% in Idaho are extremely unlikely to be replicated in Oregon. Those states used caucuses to assign delegates to the candidates, and Sanders has tended to do very well in caucuses. Both of them were also open, allowing non-Democrats to participate in the process. Oregon does neither. Here we have a closed primary, where only registered Democrats will be able to vote in the process. Then-Senator Obama only won Oregon by about 59.1-40.9% over then-Senator Clinton in 2008, and it is likely that the result will be within several points of that this time. That is why below I will analyze the 2008 results, consider the polling that has been done so far, and reflect on what they mean for the outcome of the Oregon primary this year.
In 2008 Senator Obama won by about 18 percentage points, netting only 10 delegates in the state. His path to victory included winning many smaller counties, but also all 7 of the most populous counties in the state, and mostly by double digits. He won Multnomah County by nearly 2-1, which is where most of Portland is, while the principal suburban counties, Washington County and Clackamas County, he won by nearly 19 percentage points and 9 points, respectively. He won Marion County, home of Salem, by about 8.5 points. He won Lane County, home of Eugene (and UO), by over 25 percentage points. In Central Oregon he won Deschutes County, home of Bend, by over 22 percentage points, and in Southern Oregon he won Jackson County, home of Medford and Ashland, by over 20 points. Overall he won these populous counties by slightly better than 61-39%, netting over 100,000 votes from them. He won the rest of the state by a narrower 52.7-47.3%, netting less than 10,000 votes from the rest. Nearly 77% of his votes came from those 7 counties, while about 70% of Clinton’s did.
He also carried Benton County, home of Corvallis (and OSU) by a crushing 69-30%, and Hood River County by over 20 points. His other smaller county victories were by less impressive margins. He tended to do better in areas like this, which historically voted more Republican but have trended more Democratic in recent years. Clinton only won less populous counties, mostly outside of the metropolitan areas. Her areas were older, more rural, and in some cases more Hispanic. Her only victory in the Willamette Valley was in Linn County, where the biggest city (Albany) was estimated at less than 49,000 people in 2008. Not a single city in any other county she won was even 25,000 people in 2008.
Clinton did better in areas that used to vote more Democratic but have trended against us in recent years, such as Coos and Douglas counties in southwestern Oregon, Columbia County northwest of Portland, Linn County, and much of eastern Oregon. Clinton did better than her statewide result in among the counties with the oldest populations, such as Wheeler County in Eastern Oregon, which she won, as well as Baker County, which she only narrowly lost. She won or narrowly lost every county predominantly on the Coast, as well, which are older than average in the state. She also tended to do better in counties with large Hispanic/Latino populations, such as Morrow, the only county which she won by over 20 points, and which is over 30% Hispanic/Latino. She also won Umatilla and Malheur counties, which are nearly a quarter and over 30% Hispanic/Latino according to the 2010 census. She lost Marion and several other counties with significant Hispanic/Latino populations, but by significantly less than her statewide average. In fact the only counties where Barack Obama exceeded his statewide average that have significant Hispanic/Latino populations are Washington, Multnomah, and Hood River counties, and precinct analysis shows that he tended to do better in the less Hispanic/Latino areas of those counties.
Hillary tended to do better in more Republican areas, which in a closed primary is obviously not helpful. In fact, only one county she carried was carried by President Obama in 2008 and 2012, Columbia County. This points at an issue, that counties are not counted equally in elections. I have just looked at them to try to get a feel for the candidates’ demographic strengths and weaknesses, but counties with more people, and more Democratic voters, count for more. That was a problem for Hillary in 2008, but could be better for her in 2012.
Today Democratic primaries are won or lost in the Portland metro area, which provides half of the votes in Democratic primaries. Especially in a year when Portland has an open mayoral election, like in 2012, Portland will be turning out votes, and in the 2012 primary Multnomah County alone cast about 30% of Democratic primary votes statewide. Washington and Clackamas counties will also put up a lot of votes as the 2nd and third most populous counties in the state, and both with more registered Democrats than Republicans. The Upper Valley region, which includes Lane, Benton and Linn counties, cast 15-17% of the votes in Democratic primaries, and this was Clinton’s weakest region where she lost by a bit worse than 38-62%. The Mid Valley, dominated by Salem and including Marion, Polk, and Yamhill counties, cast only 9% of the votes in the 2012 Democratic primary, while the 5 counties predominantly on the coast (Clatsop, Coos, Curry, Lincoln, Tillamook) cast only about 6%. The Southern Oregon counties together (Douglas, Jackson, and Josephine) cast less than 9% of the 2012 Democratic primary votes, as did Eastern Oregon (about 8.5%). Considering the coastal counties (where she lost 49.2-50.8%), the Mid Valley (where she got 45.7%), Southern Oregon (45.6%) and Eastern Oregon (44.8%) were her stronger regions, that was not good for her.
Why would I think that this distribution of votes, primarily in the Portland area and Willamette Valley, benefits Hillary this year when it hurt her in 2008? For several reasons, but partially because polling shows that she is doing well in the Portland area this cycle, and less well outside of it.
Very few public polls have been done in Oregon this cycle so far, and perhaps the most recent released was by local firm DHM. It was released in August of 2015, so obviously much has changed since then, but it showed Clinton narrowly leading Sanders 44-39% overall. It showed that older and female voters preferred her, younger and male voters preferred him, but the most pertinent result for our analysis here is this:
Democratic respondents who live in the tricounty area show more support for Clinton (51 percent) than those who live in the Willamette Valley (46 percent) and the rest of the state (35 percent).
“Tricounty” means Multnomah, Washington, and Clackamas counties, the principal counties of the Portland metro area, the three most populous counties in the state, and the three counties where over half of Democratic primary voters live. Another roughly quarter of Democratic primary voters live in the rest of the Willamette Valley, leaving less than a quarter in the rest of the state.
A more recent poll of just Portland voters backed Clinton’s strength there up. While Clinton led Sanders among all voters by just 37-36%, among “Democrats” and “Strong Democrats” Clinton led 57-36% and 54-43%, respectively. This shows how the closed primary helps Clinton, as while other categories the candidates basically split, among Democratic-leaning “Independent” voters Sanders leads 65-27%.
For several reasons I don’t necessarily buy these poll results, but I do buy Clinton performing better in the Portland area and worse in the rest of the state as compared to the 2008 primary. I will explore in a more detailed manner below, but the reasons are because of the demographic strengths of the different candidates in this primary cycle as compared to 2008. In 2008 Barack Obama dominated in many wealthy suburbs and other wealthy white areas, and Clinton is doing much better in those areas this year. As a result, I would expect Clinton to do relatively better in wealthier areas of Portland and the suburbs, such as Lake Oswego. President Obama carried African American voters in the 2008 primary versus Clinton, but she has cleaned up among black voters this year. Because Multnomah County has the biggest black population in the state, and roughly 75% of black voters in Oregon live in Multnomah and Washington counties, we should expect a shift towards Clinton as a result of that.
Clinton did better in more Hispanic/Latino areas in 2008, and has so far generally this year. She also did better, though, in many rural, and working class white areas where she is doing less well this year. Sanders as a result of that is likely to do relatively better than Obama did in 2008 in many of the rural counties and working class areas where Clinton did well.
The major counter to that is that Sanders, like Obama, has done better among younger voters, while Clinton has done better with older voters. Multnomah, Hood River, and Washington counties have large populations of people in their 20s and 30s, while counties where Clinton did relatively better in 2008 are generally much older on average. That may mitigate Sanders’ strength in whiter rural areas, and blunt Clinton’s in the less white and more urban areas. That factor, though, could also be mitigated by the fact that younger voters, including both non-Hispanic white and Hispanic/Latino, are more likely to register as unaffiliated than older voters, and so are likely to be less powerful a force in the closed primary than they could be otherwise.
A note on the maps to follow. There are three different kinds of maps. The first is one of three maps of the biggest cities: Portland, Eugene, and Salem. I’ve tried to make them just show the precincts of the cities. They are divided into regions of the cities showing how well Barack Obama and Hillary Clinton did in those parts of the cities. The darker purple, the stronger Obama was, the more orange, the better Clinton did, and where it is difficult to tell, the candidates were quite close to each other.
The second category of maps show counties or region of the state, where each city is colored solidly representing how the candidates did citywide, with rural and unincorporated areas colored either as the candidates did in those areas countywide, or in the larger counties in places where old and new precinct lines were easier to line up, dividing up rural regions and unincorporated communities so that they could be shown with more distinction. For example, in Washington County unincorporated communities Aloha and Bethany were considered separately from other unincorporated areas, and in Clackamas County rural precincts west of the Willamette River are separated from those south and east.
I’ve also included two maps using the old state house districts to illustrate the divides in the Portland and Salem areas in a different way. They show how the two presidential candidates did in legislative district relative to each other.
Multnomah County/Portland
Looking at more specific communities, in Portland itself Clinton lost just a bit worse than 32-68%. She did a bit better than that in the non-Multnomah County parts of Portland, getting about 33% in Portland’s Washington County precincts, and 41.5% in Clackamas County’s. In comparison to her statewide loss of about 40.9-59.1%, it looks like she lost a lot of ground by losing Portland so badly, though in the few Clackamas County precincts she barely exceeded her statewide performance.
She did particularly badly in Portland’s Northwest section in Multnomah County, with barely over 25% of the vote there, and at just above 30% she didn’t do much better in Southwest. These areas are whiter and more upscale, and so I would guess that this year she will win over some of the upscale voters she lost in 2008.
She did even worse in parts of inner Northeast Portland, though. In state House district 43 (at the time, it has since changed slightly in redistricting) lost about 20-80%. While heavily Democratic and liberal, and still majority white, the district also has a larger African American population than any other in the state. Looking just at some of the more heavily black precincts in the city, she only got about 22% of the vote, and that includes some precincts with many Hispanic/Latino voters among whom she tended to do well. As such, I would expect those area to not vote as heavily against Clinton this time around, as she is performing much better among African Americans. Since most of the African Americans in Oregon live in Multnomah County, and about 75% live in Multnomah and Washington counties, the swing among them will be highly localized.
Inner Southeast Portland, represented by state Senate district 21, was a bit worse for Clinton than the rest of Portland. She barely got over 28% of the vote there. That’s a heavily white area, known to some political insiders as “the Kremlin” because it is a stronghold of progressives in the state, a big source of Democratic votes.
Looking at the more Hispanic/Latino and Asian American areas, there were better numbers for Clinton. The most heavily Asian precinct in Portland, by my estimation, gave Clinton 48.4% to Obama’s 51.6%, much better than most of Portland. Areas that are heavily Asian in Washington County tended to vote strongly for Obama, but this is likely partly because those tend to be wealthier areas compared to the more Asian parts of Portland (this specific precinct is along 82nd Ave.) and because they are populated with different kinds of Asian voters. Asians in Portland are more of East Asian heritage, while there are higher numbers of South and Southwest Asians in Washington County. Clinton also got over 43% of the vote in some of the more Hispanic/Latino precincts in the county, which could be a bit dampened because some also had many African American voters.
Outer east Portland, represented by Senate district 24, was her strongest area in Portland. In that district she received over 51% of the vote. It makes sense, since she did better in lower income than higher income precincts. She got barely 31% of the vote in a selection of higher income precincts east of the Willamette River, but nearly 40% in lower income precincts. Outer east Portland is more working class, less white, more Hispanic/Latino and Asian American, but also with a growing black population as that community is moving out of North/Northeast Portland increasingly. But it also is less Democratic and has lower turnout than most of the rest of the city, so it is a relatively small source of votes. Will the working class whites in outer east areas swing to Sanders, or will outer east vote for the more moderate candidate, as it usually does in local races? I’m curious to see.
Maywood Park, an enclave within northeast Portland, voted closer to the statewide vote, giving Clinton about a 39-61% loss. The suburbs east of Portland, in Multnomah County, went more strongly for Clinton in 2008 than Portland did. She carried the two biggest, Gresham and Troutdale, by about 52-48%. She narrowly lost the smaller city of Fairview by only 3 points, and carried Wood Village by 58-42%. Gresham was the big prize there, though, as it is the most populous suburb and the 4th biggest city in the state after Portland, Salem, and Eugene. While Clinton is likely to still do well in East County, I think Sanders is likely to carry the county as a whole.
Washington County
In Washington County Clinton tended to do better in the further west areas less connected to the Portland area, while Obama dominated in most of the inner suburban areas in the east. The two biggest communities by far are Beaverton and Hillsboro. Beaverton is among the most Democratic suburbs in the area, and is just west of Portland. It has significant non-white populations of all kinds, but especially Hispanic/Latino and Asian American. Hillsboro is a bit larger, but less Democratic. It is the county seat in the middle of the county, and is more of an economic hub of its own, not just another suburb in the metro area. It is more heavily Hispanic/Latino, especially in the southwest of the city. While Barack Obama crushed Hillary Clinton in Beaverton by better than 61-39%, in Hillsboro she held him to a more modest 57-43% victory. In the two more heavily Hispanic/Latino precincts in southwest Hillsboro she received over 49% of the vote.
Overall in a selection of more Hispanic/Latino precincts in the county Clinton got over 45% of the vote, while in more heavily Asian American precincts she only received less than 37%. A similar divide exists if one looks at income. Among some of the highest income precincts Clinton received less than 38% of the vote, while she received over 45% in some of the lowest income.
North of Beaverton, north of Highway 26, there is a large area of unincorporated suburban neighborhoods, including heavily Asian American neighborhood of Bethany. In this area Hillary lost about 36-64%. This is the kind of upper income suburban area that I would expect her to do better in this time. Another unincorporated suburban area, Aloha, which sits southwest of Beaverton and southeast of Hillsboro, she only lost about 43-57%. Aloha has a significant Hispanic/Latino population, and is pretty working class, but has a higher rate of home ownership and far fewer apartments than the neighboring cities.
The smaller communities repeat this divide between the east and west. Clinton only slightly outperformed her statewide numbers in Tigard, a city of about 50,000, with a loss of about 41.8-58.2%, while in Tualatin and Durham she received just under 39% and 38% respectively. Just outside of Tigard and Tualatin, Hillary Clinton won the former over 55 community of King City by about 56.4-43.6% of the vote. Further southwest in Sherwood, on the county border, she lost about 42-58%. Further west of Hillsboro Hillary lost Forest Grove, a college town with a significant Hispanic/Latino population 42.8-57.2%, but won neighboring Cornelius, which is majority Hispanic/Latino, 51.8-48.2%.
Further out are the small more rural communities, where Hillary did better than she did statewide. While in the unincorporated and rural precincts combined she lost by a bit worse than 40-60, in just the rural precincts she only lost about 44-56%. In the small outlying country cities she did even better, losing Banks just 48.4-51.6%, Gaston 49.2-50.8%, and winning North Plains by about 54.1-45.9%. Given how this year’s primaries have played out, I’d expect her to fare a bit less well in these outlying areas, though she’s likely to perform better in the suburbs, where the vast majority of voters are. I still think Sanders is likely to carry the county, though.
Clackamas County
Clackamas County is older and whiter than Washington and Multnomah counties, but has fewer people and fewer Democratic primary voters. Like with Washington County there are suburbs closer to Portland and Barack Obama tended to do better towards the north and west, closer to Portland, while Hillary Clinton tended to do better further south and east, further from Portland. However she also did quite well in the more working class and less white suburban areas from Milwaukie down to Oregon City, while she did worse in the wealthier suburbs west of the Willamette River, mainly West Linn and Lake Oswego.
West of the Willamette are the generally whiter and wealthier suburbs of Lake Oswego, West Linn, and Wilsonville. Lake Oswego is at the Portland/Multnomah County border, West Linn to its south, and southwest of West Linn is Wilsonville. The further away from Portland, the better Hillary did. She lost Lake Oswego by a punishing 32.7-67.3%, West Linn by 36.7-63.3%, and Wilsonville by about 39..9-60.1%. Even the rural precincts west of the river voted for him strongly. Again, these are areas where I would expect Hillary to do better this year among the generally older and wealthier voters. As Lake Oswego is the biggest city in the county, the second most Democratic, and the voters there have among the best turnout in the state, it is perhaps the most significant source of votes in the county, and together with Wilsonville and West Linn this region has a disproportionately large share of Democratic primary votes in the county.
In the rest of the county I am less certain that she will even meet her 2008 performance. Aside from the wealthier city of Happy Valley, the rest is often more working class and white, and many voters are those who stereotypically should be with Sanders more.
South of Wilsonville is Canby and rural area, and further southeast is Molalla. While Canby is just outside the Portland area, Molalla is pretty clearly a rural community. Clinton lost Canby narrowly only 48.6-51.4%, while she carried Molalla 56.2-43.8%. These are solidly Republican-leaning communities, though, among the most in the county, and so have relatively few Democratic voters.
North of Molalla and northeast of Canby, past a few rural precincts, is historic Oregon City, the county seat. It is the second largest incorporated city in the county after Lake Oswego, and more working class. Clinton narrowly won Oregon City 50.6-49.4%, far exceeding her statewide performance by nearly 10 points. A bit east of Oregon City, heading into the foothills of the Cascade mountains, is the small rural city of Estacada. Despite being small, heavily white, and relatively deep into the hills, Estacada narrowly voted for Barack Obama in both 2008 and 2012, but Hillary Clinton widely carried it in the 2008 primary by 60.4-39.6%. Rounding out the outlying cities, Hillary Clinton won Sandy in the northeast by a narrow margin, with less than 51% of the vote.
West of Sandy are the suburbs of Happy Valley and Damascus, the most Republican suburbs in the Portland area. Hillary Clinton narrowly won the less dense Damascus area, but only slightly outperformed her statewide numbers in Happy Valley, losing 42.4-57.6%. I’d expect Clinton to do better in Happy Valley this time, as it is a wealthy community.
Between Happy Valley and the Willamette River to the west are three small working class cities and a bunch of unincorporated territory. Milwaukie borders Portland and is the most Democratic city in the county. Clinton lost it about 47.3-52.7%. South of Milwaukie, closer to Oregon City, are Gladstone and the tiny Johnson City. Clinton won Gladstone by almost exactly 4 points and dominated in Johnson City 62.7-37.3%. In the two precincts among the lowest in income and highest in population in this area, she won 57.5-42.5%.
While Hillary lost the rural and unincorporated areas by about 3 points in Clackamas County, she lost just the rural precincts by a slightly narrower 2 points. If you consider just the rural areas south and east of the Willamette River, though, it looks like she may have won them. As well, she won a number of precincts in unincorporated territory, though she appears to have narrowly lost several of the unincorporated communities like Jennings Lodge and Oak Grove.
Overall I think Clackamas County will move toward Clinton because the wealthier suburbs that cast a lot of votes will swing somewhat towards her, but it is a heavily white county, and her strong areas were more rural and working class areas where she should not be as strong this time. That could mitigate her strength this time, but I think she could carry the county given how close it was last time.
Marion County/Salem/Polk County
Marion County is the home of most of Salem, one of the state’s biggest cities, along with smaller cities Keizer, Woodburn, and numerous even smaller communities. The county’s population has a good number of younger, Hispanic/Latino voters, and older white voters. Many of the cities like Hubbard and Salem have significant Hispanic populations, while two, Woodburn and Gervais, are majority Hispanic.
Clinton only barely outperformed her statewide percentage in Salem, losing 41.6-58.4% in the Marion County portion of the city (a small portion is in Polk County). She did particularly poorly in wealthier, whiter South Salem, losing by over 20 points, and in the more central areas of Salem and downtown. She did better in areas north of Market Street and east of Highway 99/I-5, where she lost by only single digits. Those areas tend to be more Hispanic. She won, though, the unincorporated areas tucked into Salem’s east side, Hayesville and Four Corners, by a small amount. Salem’s principal suburb, Keizer, was narrowly lost by Clinton 47.7-52.3%. While working class white voters may swing more to Sanders this year, I would expect the wealthier voters in South Salem and maybe also parts of Keizer to swing more towards Clinton.
Clinton did slightly better in West Salem in Polk County, where she lost about 43-57%. That brings her citywide loss to about 41.9-58.1%. The distinction is clearer in the map using state house districts below. Clinton clearly did better in the Woodburn/North Salem district which is majority Hispanic/Latino, while she did significantly worse in the district including the college town of Monmouth and West and South Salem.
Of the small cities in North Marion, St. Paul was a tie, Donald went for Clinton 61-50 (those area actual votes), Aurora went 77-65 for Barack Obama, and Hubbard went a bit better than 53-47% for Clinton. The two majority Hispanic/Latino cities between there are Salem, Woodburn and Gervais, went for Clinton by just under 8 and 10 points, respectively.
The communities east of there, closer to the Clackamas County border, went for President Obama. He held Clinton to 42.1% in Scotts Mills, 40.8% in Mount Angel, and 37.6% in Silverton. Even further southeast, in the Santiam River canyon, Clinton won or tied all of the small cities there (Detroit, Idanha, etc.), but Obama’s margin in rural precinct 981 overwhelmed those wins, so overall he won the region by 5 votes.
On the way back west towards Salem, Clinton won the city of Sublimity with 50.8% and narrowly lost the city of Stayton with 48.8%. The small cities south of Salem Clinton carried: Jefferson with just 50.6% of the vote, Aumsville with about 55.4%, and Turner with about 58.7% of the vote.
In Polk County to the west, Clinton appears to have won the county seat of Dallas by one vote. She lost the college town, Monmouth (home of Western Oregon U) by slightly worse than 40-60%, but lost it’s neighbor Independence, which is a bit more Hispanic/Latino, by just slightly better than 45-55%. Overall she lost the rural vote in the county by about 44-46%, but won the two precincts in the rural west of the county that tend to vote for Democrats in general elections, while Obama won the more Republican rural areas in the center and east of the county.
While Sanders should do well in the working class and rural areas, Clinton will hold on to some strength in the more Hispanic/Latino areas, among older voters, and should do better in the more upper income suburban-ish precincts of South Salem and Keizer. I think she may carry the county.
Lane County
Lane County is dominated by Eugene and it’s main suburb, Springfield, but it has many smaller communities, and in most of them and the rural areas as well, there are lots of Democrats. In 2008 Barack Obama nearly swept Hillary Clinton in the cities, and I expect Bernie Sanders will similarly dominate the heavily white county, the majority of the population of which is in a big college town. In 2008 Obama crushed Clinton 69.6-30.4% in Eugene, which cast more than half of the county’s Democratic primary votes. The Springfield vote was much closer, with Clinton losing only 46.3-53.7%., but it cast only about 1/5th as many votes as Eugene. The rural and unincorporated areas of the county cast a little more than half as many votes as Eugene, and Obama swamped Clinton in those areas as well, 57.4-42.6%. No other city in the county cast even 1800 votes in the primary.
Clinton’s only wins in the county were on the extremes, winning Oakridge in the far southeast 55.4-44.6%, and Dunes City in the southwest nearly 52-48. Florence, just north of Dunes City on the Coast, she narrowly lost about 48.5-51.5%. Barack Obama carried all of the cities closer to Eugene, Junction City by only about 10 votes (50.8-49.2%), Veneta by about 55-45%, Cottage Grove about 51.3-48.7%, Creswell by about 53.8-46.2%, and Coburg about 57.7-42.3%.
Looking at Eugene with a little more detail, it’s clear where Clinton was crushed and where she has some potential. Her strongest region was West Eugene, which is a bit more working class, a little more Hispanic/Latino (for a heavily, heavily white city), and somewhat more conservative, and in that area she still lost by nearly 20 points. North Eugene is a wealthier area where she probably has room to grow, and she lost it by nearly 30 points to Barack Obama. The areas where she was really crushed, and where she is almost certain to be crushed again this year, were in the downtown/central areas, and in the UO campus/South Eugene parts of the city. Downtown/central she lost by nearly 50 points, getting under 26% of the vote. In South Eugene and the UO campus area, she received only about 22.6% of the vote. Parts of those areas are also wealthier, but those voters are strongly progressive, where the North Eugene voters are more moderate. South Eugene is the most progressive/Democratic part of the city, and that area cast thousands of more votes in the 2008 primary than any other part of the city.
Clinton did slightly worse than her average statewide but better than her average in Eugene in the unincorporated Santa Clara/River Road precincts in the northwest of Eugene, where she lost by barely more than 40-60%.
We can also take a closer look at Springfield, and see that her weakest part of the city was unsurprisingly the west end, which is where many students live. She only narrowly won the central part of the city, and lost the eastern end, which tends to be wealthier, by over 8 points. She did very well in the unincorporated precincts in and around Springfield, tending to win them by narrow margins.
While Clinton could pick up some in North Eugene and eastern Springfield if she really does better with wealthier voters, she’d have to improve more in the other areas too to really move the needle much. That also doesn’t take into account potential losses from working class and rural voters. Sanders will almost certainly dominate Lane County as Barack Obama did 8 years ago, but we also shouldn’t ignore that there are also a lot of older voters in Lane County; it’s not just a college campus, and is not dominated by one as much as Benton County to the north is by OSU. So long as older Democrats stick with Clinton she’ll get a respectable share of the vote.
Benton County
Looking briefly at Benton County, home of Corvallis and OSU, it should be a strong Sanders county. Clinton only won a few precincts there, including the small city of Monroe near Lane County, while Barack Obama crushed her in almost every other precinct, especially in Corvallis. Her only stronger area of Corvallis was in precinct 9 in the northeast of the city, but that was still only a 41-59% loss for her. In most of the rest of the city she failed to break 30% of the vote. In the precinct with the OSU campus she received less than 24% of the vote.
Overall she lost Corvallis with just about 27.8% of the vote, just a bit worse than she did in the whole county, about 30.4%. She only did marginally better in Philomath, just southwest of Corvallis, at about 33.1% of the vote. She did better in the northern part of Albany, which reaches into the county as well (most of Albany is in Linn County, which Clinton won), and Adair Village, getting 41-43% in both, and won Monroe by almost exactly points. She lost the rural/unincorporated precincts overall, though, with just 32.8% of the vote.
The issue with Benton County and Corvallis as opposed to Lane County and Eugene, is that they have college campuses of similar size, but Corvallis is much more dominated by OSU and Benton County by Corvallis than Eugene is by UO and Lane County is by Eugene. Eugene is a city of nearly three times the population of Corvallis. Corvallis cast over 63% of the votes in the county in the 2008 primary. A candidate who dominates both campuses will have a much bigger influence on Corvallis and Benton County than Eugene and Lane County. If Sanders continues to dominate with younger voters, particularly college students, he should crush Clinton in Benton County. She could have a bit of room to grow in some of the wealthier precincts, such as in the northwest of Corvallis. It’s unlikely that she can move the needle much, though, and I wouldn’t be surprised to see Sanders actually outperform Obama in Benton County so long as he contests the state.
Jackson County
Jackson County includes the most heavily Democratic city in the state, Ashland, but Ashland is dwarfed in size by Medford, one of the larger cities in the state. Ashland and the rest of the southern end of the county is where much of the Democratic strength lies, while Medford and the smaller Jacksonville are solidly Republican-leaning, and Central Point and other areas north of Medford in the county are even more Republican.
Barack Obama carried Jackson County by about 60-39%, only slightly wider than his statewide margin. That outperformance is definitely a result of how he won overwhelmingly in the southern end of the county, winning much of Ashland and the surrounding region by 3-1 or better. In fact he defeated her 75.9-24.1% in Ashland itself. Ashland is heavily white, wealthier than average, and it’s a college town with Southern Oregon University. Obama won the cities of Talent and Phoenix between Ashland and Medford, but did less well further north in Phoenix (54.7%) than in Talent (61%).
Hillary Clinton did better in Medford, earning 44.4% of the vote. She did less well in the whiter, wealthier south and east of the city, but nearly won the western, more Hispanic/Latino and more working class and downtown areas of the city. That is to an extent in contrast to her performance in cities like Eugene and Portland, where she tended to do better in less Democratic parts of those cities, but in Medford among her best precincts were the most Democratic. Jacksonville just to the west, which is much smaller, gave her only 36.3% of the vote.
Hillary Clinton won more in the north of the county, with narrow losses in Central Point (49.7-50.3%) and Eagle Point (47.7-52.3%), but winning or tying all of the other cities. She carried Rogue River 56.4-43.6%, and tied in Gold Hill, the cities west of the Medford area. Clinton won Shady Cove in the far north 56.9-43.1%, and Butte Falls in the east 71-61 (in raw votes). The unincorporated and more heavily Hispanic/Latino White City north of Medford Clinton carried 55.5-44.%. While Clinton lost the rural/unincorporated vote in the county 41.7-58.3% overall, she did better in the north of the county than the south.
While I expect Sanders will still do well among Ashland voters, and is likely to do better than Obama in rural areas, I think Clinton will do better in higher income precincts, and potentially even among higher income voters in Ashland.
Deschutes County
Clinton lost Deschutes County in Central Oregon by just a bit more than Jackson County, closer to 38-61%. That was because she was crushed in Bend, the biggest city east of the Cascades and the population center in Deschutes County, 32.6-67.4%. In the resort community of Sunriver, she only received 29.1% of the vote. In the rest of the county, though, she did much better.
Looking at Bend with a bit more nuance, though, she did worse on the west side of the city, west of the Deschutes River and Highway 97, that is, than on the east. The west side is a bit whiter (though the whole city is very, very white) and wealthier, with the only lower income neighborhoods being on the east side. Clinton received only about 23.6% of the vote on the west side. In the only precinct with a large Hispanic/Latino population (or large nonwhite population at all) she got over 38% of the vote, significantly better than the rest of the city. In that and the other lower income precincts she did several points better than in the rest of the city as well.
I still do expect Sanders to win strongly in Bend as Obama did, but given how Clinton is performing better with urban and suburban wealthier white voters than in 2008, I expect her to do better in Bend this time around. In contrast I expect her to perform relatively less well in the rest of the county. In 2008 she lost the rural/unincorporated vote 43.3-56.7%, significantly better than she did countywide. While she lost the tourist town of Sisters by 38.5-61.5%, she did much better in the other cities in the county, apparently losing Redmond by only 4 votes (49.9-50.1%), but winning Terrebonne’s precinct narrowly and carrying La Pine widely, 68.7-31.3%. Bend really decides the Democratic primary, though, as it cast over 11,200 votes for Clinton and Obama in 2008, compared to about 2,400 in Redmond, and less than 8,300 in the rest of the county. The candidate winning Bend in the primary is likely to win the county, and at this point it looks to be a Sanders county.
Jefferson County
Jefferson County is just north of Deschutes County, and is much smaller in population, but more diverse. It is the least white county in the state, with substantial Hispanic/Latino and Native American populations. Barack Obama defeated Hillary Clinton 51.7-48.3%, very narrow compared to his statewide victory. His victory was accomplished through winning most of the west and east ends of the county, while Clinton won the more Hispanic/Latino county seat of Madras. He defeated her soundly in western end, carrying Warm Springs, the Native American community, by slightly better than 60-40%, and Camp Sherman by 75.4-24.6%. The suburb of Madras, Metolius, was Clinton’s strongest precinct, where she claimed over 63% of the vote (coincidentally the small community only cast 63 votes between her and her opponent). She only narrowly carried the larger cities of Culver and Madras by about 51-49% and 51.6-48.4%, respectively.
This county has had the largest increase in voter registration since the implementation of automatic registration of any county in the state. I’m interested in seeing how that affects the outcome of elections there, if it will at all. Aside from that, I would expect Sanders to do somewhat better than Obama did among white voters there, while I think Clinton will likely hold on to many of her older and Hispanic/Latino voters, but probably lose a little more ground overall.
Hood River County
Hood River County is a smaller one in terms of population, and is just east of Multnomah County but not really considered a part of the Portland metro area. It is more of the gateway to the Columbia River Gorge and Eastern Oregon. There is a sizeable Hispanic/Latino population, and also a large young white professional population. It is one of the historically Republican but now liberal counties that voted strongly for Barack Obama in 2008 in the primary, and was among his strongest counties in his general election victories too.
In the 2008 primary Clinton lost by just over 20 points countywide, and by even more in the city of Hood River. Of the precincts in the city, Clinton’s worst was one of the more heavily Hispanic/Latino ones, but also her best was. She actually did better in the rural/unincorporated precincts and in the other, smaller city in the county, Cascade Locks. Surprisingly the city of Hood River, her weakest area, was also lower income than the rest of the county, so the income pattern in voting was reversed here. She lost the city of Hood River about 35-65%, which cast about 1/3 of the county’s primary votes, and won the much smaller Cascade Locks 57.7-42.3%, though it cast less than 5% of the county’s votes. She also barely won Odell’s precinct by 3 votes out of about 250.
I am hesitant to make a prediction about how things will change this year, because Clinton already did better in areas I would expect her to now.
Douglas County
Douglas County was the second biggest that Clinton won. The heavily white, more rural, and historically more Democratic county voted for Clinton 51.3-48.7%. Although heavily Republican in most general elections, it typically gives populist Congressman Peter DeFazio (D-Springfield) much better treatment. He often runs 20 points or more higher than other Democrats on the ballot. Demographics seem to point to Sanders doing better in that kind of county this time.
Despite winning countywide, Clinton narrowly lost the biggest city, Roseburg, and some of the communities closer to Eugene in the north of the county. President Obama carried Roseburg about 53-47%, and the small northern cities Oakland, Elkton, and Drain by margins of about 57-43%, 60-40%, and 50.8-49.2%, respectively (Elkton is tiny; that win in actual votes was 18-12). Clinton did better in a couple northern communities, including Reedsport (winning about 55-45%) on the coast, and Sutherlin, the second largest city (winning about 59.8-40.2%). She nearly swept the cities south of Roseburg, losing only Riddle 47.6-52.4%. She won Canyonville better than 60-40%, Myrtle Creek 58-42%, Winston nearly 56-44%, and Glendale by actual votes 43-37.
I would look for a similar outcome this time, with Sanders performing better in the north than the south. I think this is one of his best shots to flip a county from Clinton, though.
Klamath County
Klamath County can be defined as either part of Southern or Eastern Oregon, and is a solidly Republican County. While generally very white, and lower income, it also has a Native American presence and a somewhat significant presence of nonwhites, particularly Hispanics/Latinos, particularly in Klamath Falls and some of the smaller cities. A substantial unincorporated suburb, Altamont, is roughly the same size as Klamath Falls, adjacent to it, and somewhat whiter and more well-to-do than Klamath Falls. Outside of Klamath Falls and Altamont the county is rural and sparsely populated.
Clinton carried the county by about 54-46% in 2008, making it one of her strongest in the state. Only about 6,500 voters cast ballots in the Democratic primary that year, which was still over 70% turnout for Democrats in this relatively low population (though fairly high for southeast Oregon), strongly Republican county. Clinton carried every city in the county except one, she carried the unincorporated community of Altamont, and she carried the rural areas of the county as well, for overall a fairly consistent performance countywide.
Clinton won Klamath Falls only narrowly, 50.5-49.5%, or by about 19 of nearly 1,900 votes cast. Altamont cast slightly more votes, just over 1,900, and Clinton won that whiter, higher income community by a more pronounced, nearly 57-43%. Barack Obama did narrowly win the two precincts at the north of Klamath Falls, where the Oregon Institute of Technology is, and which are higher income than the rest of Klamath Falls, but its interesting that he still did better in the poorer, slightly more Hispanic/Latino Klamath Falls than in Altamont. That goes against type for that election. Clinton gt abut 54.4% of the vote in rural areas, slightly better than her result countywide.
Are the smaller cities of any interest? Probably not, because they cast very few votes, but I suppose they kind of are demographically. Chiloquin was nearly half Native American (49.2%) in the last census, Malin majority Hispanic/Latino, Merrill over 40% Hispanic/Latino, and Bonanza over a quarter. The low vote counts really make them statistically insignificant, though. Clinton narrowly won Chiloquin 30-28, tied in Malin 14-14, won Merrill 28-22, and Bonanza 15-13.
Given the vote comes down to Klamath Falls, Altamont, and the rest, the estimate is simpler. I think Sanders could pick this one up, and it would be mostly thanks to doing better in rural areas than Obama did. I think relative to her 2008 performance Clinton will have stronger showings in Klamath Falls and Altamont than the rural areas.
Conclusions
Assuming that both candidates are still in this race and at least one is contesting the state, I think there are more counties where Clinton will underperform her 2008 numbers than exceed them. However, the majority of the primary vote will come from the Portland metro area, and even more the whole of the Willamette Valley which includes it, and because I expect a good number of voters who supported Obama by large margins there to flip to Hillary, I think she is likely to exceed her 2008 statewide vote share overall. In 2008 then-Senator Obama dominated among African Americans and in wealthy urban and suburban areas, while this time around the economic divide is not very stark, and Clinton is the one winning African American voters. Clinton on the other hand is not doing relatively better in working class and rural areas, but is still doing better with older and Hispanic/Latino voters. Overall, Clinton will do relatively better among African Americans and in wealthier urban and suburban areas, and because of that will perform better in the big, populous counties than she did before (when she lost all of the 7 most populous counties). I think she could carry several of them, with Clackamas and Marion being the most likely. I could definitely be wrong about that, and could see Sanders slightly outperforming Obama, but only by at most a few points.
Sanders will win the same progressive strongholds that Obama did in 2008: Portland, particularly Southeast Portland, Eugene, and in particular South Eugene, Corvallis, and Ashland. He may do better than Obama did in some working class white areas, and some rural areas. He will almost certainly win the state. The best result I foresee for Hillary Clinton is around a 45-55% loss, the worst around a 40-60% loss. I expect somewhere in the middle, around a 57-43% Sanders victory.
In terms of the delegate allocation among congressional districts, it looks like Clinton’s best districts were the 2nd (44%) and the 5th (43%), while her worst was the 3rd (37%), with the 1st (40%) and 4th (41%) being middling. Given my demographic assumptions about the change from 2008, I think she should improve a bit in the 1st, 3rd, and 5th, while I think she could do worse in the 2nd, and possibly the 4th as well. I would not be surprised for the 4th to be Sanders’ best, and the 5th to be Clinton’s.