The Changing Electorate (and the implications for down-ballot races)

Cross-posted at Election Inspection 

(Note: due to formating issues, I didn't post the charts here, to see how Obama did compared to Kerry, visit the Election Inspection link) 

I've actually been quite interested in doing a comparison of how Obama did compared to the last Democratic nominee (Kerry). Here's the difference between Obama and Kerry's margins in each state (for reference, I subtracted Kerry's margin from Obama's margin to get the final number, for example, if Obama's margin in California was 24 and Kerry's margin was 9, the equation would be 24-9=15).

Obviously, since Obama won the popular vote by 7, while Kerry lost it by 3, Obama is going to outperform Kerry almost everywhere, and speaking of, the only states where Obama did not outperform Kerry were in Alaska (-1), Arkansas (-11), Louisiana (-4), Tennessee, (-1), Oklahoma (0), and West Virginia. This, however, only tells us what we already know, Obama outperformed Kerry almost everywhere. A more important question to ask would be, where did Obama do better than Kerry relative to how the entire country did (to put it another way, we know that Kerry won California by 9 points, but he lost the national popular vote by 3 points, so Kerry actually ran 12 points higher in California than in the country, and Obama, who won California by 24 points but won the popular vote nationwide by 7 points, performed 17 points better than the country at large. Subtracting Kerry's performance in California compared to the country at large from Obama's same performance means California voted 5 points more Democratic relative to to the rest of the country than it did 4 years ago).

So how did Obama do in these other states compared to the national vote relative to Kerry?

(Follow link at the top for a look at the relative performance of Obama to Kerry in each state)

This gives us a much better picture of which states, in any given year, are moving more Democratic, and which ones are stalling out. Of course, it would be smart to keep in mind that some of these numbers have to be taken in context of home state effects of presidential and vice presidential candidates (Arizona, Alaska, Illinois, Hawaii, and Delaware are the home-states of McCain, Palin, Obama, and Biden respectively; while Texas, Wyoming, Massachusetts, and North Carolina are the home-states of Bush, Cheney, Kerry, and Edwards, if some people over/underperform in certain states and regions, it has to be taken in this context). The glaring exception to the home-state advantage here is North Carolina, where Obama performed three points better relatively to his popular vote standing than Kerry did (and could easily be attributed to the growth of the Raleigh/Durham/Chapel Hill area). The states where Obama had the highest outperformance of Kerry's standing were in Hawaii (+26), Indiana (+12), North Dakota (+9), and a three-way tie between Utah, Montana, and Nebraska (+7 each). Obama's top under-performances, by comparison, were in Arkansas (-21), Louisiana (-14), Alaska (-11), Tennessee (-11), with a tie between West Virginia and Oklahoma (-10 each). There are, of course, a bunch of others, but generally speaking, we can say that by comparison, Obama generally underperformed Kerry in the south and the northeast (the exceptions being Vermont, Virginia, Georgia, Delaware, Connecticut, and North Carolina), while he generally outperformed Kerry in the midwest and the west, particularly where there was a large Hispanic population (New Mexico, Nevada, Colorado, and California). Obama seemed to stick pretty close to Kerry's relative performance in Washington State, Oregon, and Iowa (in fact, it seems that compared to the country, Iowa seems to have a consistant Democratic lean, as it changed exactly zero relative to the country)

This has extremely important ramifications for both presidential and down-ticket races in the future, for example, three states which Obama won which Kerry did not (Ohio, Florida, and Iowa) might seem to be massive improvements for the Democrats compared to how Kerry did, but in actuality, Obama underperformed Kerry relative to the rest of the country in Florida and Ohio, while Iowa stayed the same relative to the rest of the country (that is to say, in both 2004 and 2008, Iowa was roughly three points more Democratic than the country at large) (of course, for Ohio, Kerry actually did relatively better than most Democrats normally do in Ohio, but it usually tends to vote slightly more Republican than the rest of the country, whereas Ohio voted slightly LESS Republican than the national vote in 2008). Now, relatively speaking, Obama tended to GREATLY outperform Kerry in the midwest (Obama's relative performance in Wisconsin was 2 points better, in South Dakota was 3 points better, in Nebraska it was 7 points better, and a full 9 points better(!)). Of course, Obama did, relatively speaking, underperform Kerry in Minnesota, but that might be more a function of McCain spending a dispropotionate amount of time and resources in Minnesota (one of the only places where McCain was significantly outspending Obama on both field organization and advertising). The places where Obama really outperformed Kerry though were in the southwest and the mountain west (Obama outperformed Kerry by 3 points in Colorado, 4 points in Idaho, 5 points in Nevada, 5 points in California, , 6 points in New Mexico, 7 points in Montana, and 7 points in Utah. Like I said above, Obama did tend to underperform in the south, but the three places where Obama outperformed Kerry are states which have strong implications for state-wide Democrats are in Georgia (+2), North Carolina (+3), and Virginia (+4). The other two big deals are California (which has become almost as Democratic as New York) and Indiana (which went from being 18 points more Republican than the country to being only 6 points more Republican).

Democrats are probably going to have a harder time getting elected in Southern states like Alabama, Mississippi, Louisiana, Tennessee, Kentucky, and Oklahoma, but strong Democrats are going to have a much easier time running in states like Montana, Colorado, New Mexico, Indiana, North Carolina, Virginia, North Dakota, and South Dakota. Furthermore, with California's gubenatorial race in 2010, if the Democrats don't rip each other apart like they did in 2006, they should have an extremely good chance at winning the governor's mansion, and controlling redistricting for the census.

Detailed County Predictions for Presidential Race

I spent hours writing this as a Daily Kos diary and virtually nobody read it, but I figured some around here would appreciate the regional specifics and hope some enjoy the read.

It appears incredibly likely at this point that Barack Obama will be elected President in nine days, and I’m excited as hell to see the county maps roll in on CNN and USA Today websites on election night to see how much more blue there is on the national county map than in 2000 or 2004, when more than 80% of the terrain was colored red.  It seems certain that Obama will dramatically improve upon the 582 counties and independent cities that John Kerry won in 2008, but I’m wondering how many of those 582 from last time will be lost.  The ideal would be none, but looking at polling data from several states, it seems likely that a number of them are at serious risk.  Details below the fold.

Alabama–11 counties went for John Kerry in 2004, all in a narrow belt of heavily black counties south of Birmingham.  The only two that might be at risk are Montgomery County and Russell Counties, both of which Kerry won narrowly.  My bet is that higher black turnout than what was seen in 2004 will keep those counties blue, but I’m skeptical whether Obama will pick up more than those 11 Alabama counties.

Alaska–no counties so I’m scarcely interested in their undefined “election districts”

Arizona–Right now it seems Obama is poised to overperform Kerry in McCain’s home state.  Kerry won four Arizona counties, all four of which seem near slam-dunks to shade blue again.  There aren’t too many likely candidates for pickups beyond those four though.

Arkansas–Here’s where Obama is likely to cede some territory.  Obama is underperforming Kerry in Arkansas polls, and given that there are some Yellow Dog Democrat strongholds in Arkansas that are more than 95% white, it seems unlikely that Obama will hold all 21 of Kerry’s counties.  Likely gone:  rural counties in the northeast such as Clay, Randolph, Lawrence, and Poinsett; and at least a few southern counties that aren’t majority black such as Little River, Hempstead, and Bradley.  High black turnout in some of the 30+% black turnout counties could help flip one or two 2004 Bush counties to Obama, but I’d bet against it.

California–Kerry won 22 counties in California, but with few exceptions, the CA counties will remain polarized.  I’m not anticipating Obama losing any of the Kerry counties, but only see a handful of opportunities to pick off Bush counties, such as Ventura County and San Joaquin County.

Colorado–Kerry won 19 Colorado counties in 2004, several more than Gore did, picking off a number of Rocky Mountain counties but losing ground in the Hispanic-heavy region in and around Pueblo.  I’m hoping Obama regains the footing with Latinos that Kerry lost in that region and pick off Huerfano and Alamosa Counties, but more important to Obama’s statewide victory is the need to improve upon Kerry’s performance in the Denver suburbs.  He needs to grow the Kerry margin in Adams County and pick off at least two out of three of Jefferson, Arapahoe, and Larimer Counties, all of which went narrowly Bush in 2004.

Connecticut–Kerry won seven out of eight but lost Litchfield County in the northwest corner of CT.  I think Obama will win that one back as Gore did in 2000.

Delaware–With Biden on the ticket, I’m hopeful Obama can win at least one of two of Republican-leaning Kent and Sussex Counties.  Fortunately for Dems, they can win handily with New Castle County up north, the one county Kerry won in Delaware.

Florida–Kerry pulled out only 11 Florida counties.  My sense is that Obama holds those 11 (although fast-growing St. Lucie County is a question mark), and will probably pick up several more battlegrounds, ideally both Pinellas and Hillsborough (along with possibly Flagler, a Gore county) in the Tampa-St. Petersburg area.  Osceola County near Orlando is an option, as are a few rural counties near Tallahassee with high black populations that have narrowly gone Bush in the past but may benefit from higher black turnout this year.  Whether this would be enough for Obama to win Florida remains a question mark.

Georgia–Kerry won 26 Georgia counties, which may sound impressive until you realize there are 159 counties in Georgia.  High black turnout and a hard-fought Senate race seem likely to expand the number of Obama counties, but mostly in rural regions of southwestern and east-central Georgia.  It’s doubtful that any of the more heavily populated Bush counties in Georgia will turn blue this year.

Hawaii–Kerry won all four counties, but only narrowly eked out the population center of Honolulu County.  Obama should vastly overperform in all four Hawaii counties.

Idaho–Kerry won only one county, Blaine County, which is where he skiied in Sun Valley.  That will go Obama this year, and I suspect Latah County, which includes the college town of Moscow, will as well.  Beyond that, the pickings will be mighty slim for blue territory this year.

Illinois–Kerry won only 15 of Illinois’ 105 (is that correct?) counties in 2004.  Needless to say, Obama will perform remarkably better than that.  But I’m not sure exactly how many more.  Will Obama win the long-standing GOP stronghold of Du Page County next to Chicago?  It’s possible, but I wouldn’t bet money on it in a national election.  Overall, I’d bet that Obama wins slightly more than half of the Illinois counties, but expect a sea of red in the southern Illinois counties (the region south of St. Louis and Decatur, aside from a handful of Dem strongholds like Carbondale and Cairo).  Several of the southern Illinois counties went for Alan Keyes in 2004 and many more voted for Hillary Clinton in the primary.

Indiana–This one should be fun.  Kerry won only four Indiana counties, but even if we assume Obama falls a few points short of winning the state, alot more turf would turn blue since 2004.  I expect most of Indiana’s population centers would turn blue, including counties like Vigo (Terre Haute), St. Joseph (South Bend), Howard (Kokomo), Tippecanoe (Lafayette), and Delaware (Muncie).  It’s possible even hard-core conservative Allen County (Fort Wayne) could turn blue.  The wild cards that will determine if Obama wins or loses Indiana will be southern Indiana, including Vanderburgh (Evansville) and the north river Louisville counties of Clark and Floyd.

Iowa–Even when narrowly losing Iowa, Kerry still won 32 of Iowa’s 99 counties.  If Obama is ahead by double digits this year, expect him to win the “Harkin coalition” of 60-some counties, essentially everywhere but the western two tiers of counties, and a few outlying GOP bastions.  Southern Iowa thinks and votes like conservative northern Missouri, so Obama may fall short in most of the territory south of Des Moines.  Still, it seems very unlikely that any of Iowa’s 32 Kerry counties will go McCain.

Kansas–Kerry won two Kansas counties.  Those two are still solid, but Obama is unlikely to gain much ground, but could conceivably win Shawnee County (Topeka).

Kentucky–Not good.  Kerry won 12 counties out of 120 in 2004, all but one in culturally conservative eastern Kentucky coal country.  The only Kerry county I feel solid about regarding Obama’s chances is Jefferson County (Louisville), where polling internals suggest Obama is vastly overperforming Kerry and likely keeping the statewide margins in Kentucky in line with 2004.  That indicates Obama is underperforming elsewhere in the state, and the very eastern Kentucky counties that Kerry won, by more than 60% in three of those counties, are the places where Obama was most fervently destroyed in the primaries, pulling in as little as 5% in a couple of them.  That suggests a serious cultural disconnect in play and I wouldn’t be surprised if Obama lost all 11 of those east Kentucky Kerry counties.  If he wins any of them, I anticipate they would be the staunch Democratic counties of Elliott, Floyd, Knott, and Breathitt.  Obama’s only pickup opportunity in Kentucky would seem to be Fayette County (Lexington) or Franklin County (Frankfurt), but I wouldn’t bet on either of those.

Louisiana–This one’s completely up in the air due to Hurricane Katrina displacements.  There are conceivably a few heavy black Bush counties that could be picked off with a high African-American turnout, but I’m not optimistic that Obama will net much more than the 10 parishes Kerry won.

Maine–Kerry won 14 out of 16 Maine counties in 2008.  It seems tough to imagine the other two going into the Obama column this year, but I suppose Washington County would be possible.  I doubt McCain will gain any of the Kerry counties, but Penobscot County is possible given Kerry won it by a half-percentage point.

Maryland–Five of Maryland’s 23 counties, along with the independent city of Baltimore, are usually all the Democrats can hope to get in Maryland given the overwhelming Republican tendencies of the rural parts of the state.  Both Gore and Kerry won these six and these six alone.  Obama could conceivably pick off Anne Arundel County and maybe one of those soft GOP counties on the Eastern Shore like Somerset, but he won’t encroach into GOP territory too much.

Massachusetts–Kerry won all 14 Massachusetts counties, and Obama should do the same.  If any switch, it’ll be Plymouth or Barnstable Counties in the southeast, but I doubt that’ll happen.

Michigan–Kerry won only 15 out of Michigan’s 83 counties, but as is the case with most of the Upper Midwest, Obama is poised to really build upon that number and conceivably win an outright majority of those counties.  Bush won many of his Michigan counties with very soft margins, suggesting a partisan breeze of only five points in Obama’s direction will turn multiple counties blue, including populous counties such as Macomb, Monroe, and Calhoun, as well as large numbers of thinly populated rural counties in northern Michigan and on the Upper Peninsuala.

Minnesota–Kerry won 24 of Minnesota’s 87 counties, and demographics suggest Obama stands point to significantly grow upon that, again potentially winning an outright majority of Minnesota counties.  Internals from polls of North and South Dakota suggest Obama is winning the eastern farm counties of both states.  If that’s true, it likewise means Obama is winning the populist farm counties of western Minnesota, which Gore and Kerry both got smashed in.

Mississippi–Kerry won an impressive 24 counties in Mississippi, but don’t expect that to change much in 2008 simply due to the racial breakdown of those counties.  Obama is likely to grow Kerry’s margins in most of those 24 counties, most of which are majority black, but I’d be surprised if he picked off more than one or two of the 2004 Bush counties.

Missouri–Kerry won three counties and the city of St. Louis in Minnesota…out of 115 counties!!!  Obama will do better than that, but not significantly so.  There are a number of counties encircling metropolitan St. Louis that should be favorable turf.  Boone County (Columbia) seems like Obama territory, and possibly Buchanan County (St. Joseph) north of Kansas City.  Still, I’m not expecting more than 10 Missouri counties for Obama even if he wins the state.

Montana–Kerry won six Montana counties, all of which seem solid for Obama.  Beyond that, it’s not inconceivable to imagine Obama victories in populous (at least for Montana!) Yellowstone County (Billings), Cascade County (Great Falls), and Lewis and Clark County (Helena), along with rural counties like Hill and Blaine that are winnable for Democrats with high Native American turnout.

Nebraska–Kerry won one Nebraska county in 2004.  If Obama gets as many as five, he’ll have done better than any Democratic Presidential candidate in my lifetime.  Possible pickoffs:  rural Saline and Dakota Counties, Lancaster COunty (Lincoln), and conceivably but doubtful in my opinion, Douglas County (Omaha).

Nevada–Kerry won only Clark County in 2004, but narrowed the gap to four points in Washoe County (Reno).  Obama needs Washoe to win the state, and early indications are that he’s winning it, but none of the other 15 Nevada counties are likely to be in play.

New Hampshire–Kerry won six of 10 New Hampshire Counties in 2004, but not among them were the two most populous (Hillsborough and Rockingham).  With current trendlines, it seems as though Obama should win those two counties, but is still unlikely to win the other two Bush counties.

New Jersey–Kerry won 12 of New Jersey’s 21 counties, two fewer than Gore won in 2000.  Those 12 counties seem secure, and Obama could pick off the Gore county of Salem (increasingly Republican Gore County, Monmouth, seems like a stretch) but is unlikely to pick off any of the twice-Bush counties.

New Mexico–Kerry won 12 counties in New Mexico yet lost the state.  If Obama wins comfortably this time, as polls suggest he will, he’ll probably take a handful of additional counties and grow his margins in some of the soft Kerry counties, but I suspect most of the ranch counties in southern and eastern New Mexico will remain red.

New York–Kerry won only 21 New York counties, but I anticipate Obama is poised to overperform Kerry in a number of upstate New York counties (as well as win back Rockland County and Staten Island in metropolitan NYC), particularly those that went Gore in 2000.  McCain will still probably win close to half of the upstate New York counties though, particularly those out in Tom Reynolds and Randy Kuhl country.

North Carolina–Kerry won 20 out of 100 North Carolina counties in 2004, five fewer than Gore did four years earlier.  Expect Obama to win as many as 35, with several heavily black rural counties in eastern North Carolina having gone narrowly Bush in 2004 that will be easier picking this time.  Obama should dramatically grow his margins in the population centers, picking off two pretty significant prizes including Wake County (Raleigh, Cary) and Buncombe County (Asheville).

North Dakota–Kerry won only four of North Dakota’s 53 counties in 2004, but polling internals shows Obama is ahead in eastern North Dakota now.  It’s possible Obama could be the first Democratic Presidential nominee in my lifetime to win Cass County (Fargo) and Grand Forks County.  At the very least, expect Obama to win 15 or more counties in North Dakota, as several of those eastern ND sugar-growing counties should be favorable territory for him if the statewide race is as close as polls suggest.  A high native American turnout could turn the tide in a couple western North Dakota counties such as Mountrail, but for the most part, western ND will still be bright red.

Ohio–Very tough to call this one as polls are all over the place.  Kerry won only 16 Ohio counties, the same number as Gore won.  Obama has to do better than that to win the state.  A high black turnout in Cincinnati is very likely to flip Hamilton County blue and Lake County in the Cleveland suburbs seems like a decent bet to flip.  But if we’re to believe the polling median that Obama is 5-7 points ahead in Ohio, that would mean Obama is likely winning some of the southern Ohio counties like Scioto (Portsmouth) and Ross (Chillicothe) that are generally bellwethers for statewide victory in Ohio.  There’s too much conflicting information at this point, and the poll spread suggests Obama could win as few as 12-15 Ohio counties or as many as 30.

Oklahoma–Gore won nine Oklahoma counties in 2000….Kerry won zero.  Expect Obama to repeat Kerry’s performance.

Oregon–Kerry won only eight counties amongst a sea of territorial red in this blue state, but Obama’s poll leads in Oregon are so substantial than I suspect Obama will win double his number of county victories, though still mostly in the northwestern quadrant of the state.

Pennsylvania–Another state that’s difficult to call.  Kerry won only 13 counties here, ceding some territory that Gore won in western Pennsylvania.  If Obama is really winning Pennsylvania by 10 points as the polls suggest, he probably is winning all or most of those 13 Kerry counties and then some, but the conventional wisdom is that Obama is underperforming Kerry in western PA, and could end up losing the culturally conservative Kerry counties of Beaver, Washington, and Fayette.  That certainly is possible, particularly if we assume Obama’s margins in PA are entirely the product of his running up the score in suburban Philadelphia counties such as Montgomery, Delaware, and Bucks.  Tough to call, but I think we lose some Kerry counties in PA, but also pick up fast-growing Chester County in exurban Philadelphia for Obama.

Rhode Island–Kerry handily won all five Rhode Island counties.  Obama will do the same.

South Carolina–Kerry won 15 counties, most of them majority black.  Obama could win back two or three more with high black turnout, but I’m not expecting to grow the county map much within South Carolina.  The I-95 corridor will remain blue, the rest of the state will remain red.

South Dakota–Polling internals suggest Obama is ahead in northeastern South Dakota.  That means Obama is likely to improve upon the nine SD counties that went for Kerry (but keep in mind that several of the Indian reservation counties will have a lower turnout without a battleground Senate race on the ballot).  It’s conceivable Obama could win 25 counties in SD, since most of the Daschle coalition in eastern SD are smaller counties size way that tend to vote as a bloc.  I doubt Minnehaha County (Sioux Falls), now the destination for college Republicans across the country, will go blue, but expect other population centers like Brookings County (Brookings), Beadle County (Huron), and Brown County (Aberdeen) to either vote Obama outright or come very close to doing so if current polling is to be believed.

Tennessee–Gore won 36 of Tennessee’s 95 counties, Kerry halved that by 18, and if Obama is lucky, he’ll get by only halving Kerry’s numbers to nine counties this year.  Obama should hold population centers Shelby County (Memphis) and Davidson County (Nashville) along with a couple of heavily black rural counties in West Tennessee, but some of the Yellow Dog Democrat rural counties in West and Middle Tennessee that have slowly slipping away for 20 years, including Smith County, home of Al Gore’s hometown of Carthage, will probably turn red.  A handful of tiny deep blue counties such as Jackson, Houston, and Trousdale may stay blue, but I anticipate losing the majority of them.

Texas–Kerry won only 18 counties in Texas, but I’m anticipating Obama to do much better.  Several of the majority-Hispanic south Texas counties (like Cameron and Frio) went narrowly Bush in 2004, and I expect to win them back.  Several population centers in Texas could tip.  I’d bet heavily on Dallas County turning blue this year, as Kerry narrowly missed it in 2004, but beyond that, Bexar County (San Antonio) and Nueces County (Corpus Christi) are also within the realm.  We might be close in the big prize of Harris County (metropolitan Houston) but I still think Obama will fall short of victory. Overall, I’m betting on 30 or more Obama counties in Texas this year.  The bad news…he’ll still lose more than 200 of them.

Utah–Kerry didn’t win any counties in Utah, but I actually think Obama has a chance in three of them this year, including the big prize of Salt Lake County.  The youthful mountain counties of Grand and San Juan in the state’s southeast corner are more likely to flip though.

Vermont–Kerry won 13 of 14 counties in Vermont and Obama will probably win the same 13.  Tiny Essex County in the northeast corner of the state seems likely to remain red though.

Virginia–Kerry won 13 Virginia counties and 19 independent cities in Virginia, and Obama should win most if not all of those locales, with the possible exception of those two southwest Virginia coal counties which remain question marks.  The good news it that Obama is poised to win handily in the important places, namely fast-growing exurban NoVa counties Prince William and Loudoun, both of which would have been inconceivable to see turning blue in 2004.

Washington–Kerry won 12 counties in WA, all on or near the coast.  There are probably about five or six additional counties in play for Obama, again mostly in the western third of the state.  Spokane County used to be a pretty reliable Democratic county, even going for Dukakis in 1988, but seems way out-of-reach for Obama 20 years later.

West Virginia–Polls seem to have backslid a little bit in West Virginia in the last week, and I suspect the end result will be just as bad as Bush’s 13-point victory in the state in 2004.  My anticipation is that the county map will look a little different though.  Obama may likely pick off 2004 Bush counties in population centers like Kanawha County (Charleston) and Monongalia County (Morgantown), but I’m not confident he’ll hold too many of the southern WV coal counties that Kerry won, aside from possibly McDowell and Boone, the bluest two counties down there.  Overall, I’d be surprised if Obama got more than seven WV counties, compared to Kerry’s nine.

Wisconsin–Even though Kerry won by the skin of his teeth, he still managed to win 27 of Wisconsin’s 72 politically polarized counties.  There are probably about 15 more counties in play if Obama has a double-digit lead in Wisconsin as polls indicate.  Most of the Obama pickups are likely to be found in northwestern Wisconsin.

Wyoming–Kerry won one Wyoming County, Teton, and Obama will probably win that and that alone as well.  Albany County (Laramie) may have enough youth votes to turn blue.