This is Bill.
Bill is on the Internet looking at election results in the early phase of the election.
Bill sees Donald Trump leading in key swing states.
Bill freaks out and panics!
Bill does not know that small rural counties tend to report first and also tend to be more Republican.
Don’t be like Bill.
Instead, use our county benchmarks! This is a tradition for every general election at Daily Kos Elections, which breaks the key states down into what counties actually matter (so that you don’t have to break out the road atlas and stare at the fine print in the back until you find “Geez, the Republican has 80 percent of the vote in McGillicudy County, but there are only 5,000 people there? Who cares!”). “Actually matter,” for our purpose, means counties where the 2012 vote total is 2 or more percent of the state’s 2012 vote total.
Even more importantly, it gives you estimates for what the Democratic candidate should be earning in each of those key counties, in order to be on track statewide. Different counties might come in at different times of the night, but this way you can keep track of whether the early vote in the key counties is on target for a victory.
One caveat: this is naive extrapolation at its finest. It assumes uniform swing from county to county based on the 2012 results. In some counties there may be turnout surges where they represent a bigger percentage of the statewide total than they did in 2012 (which, based on early voting reports, seems to be happening in Hispanic-heavy counties in Sun Belt states). The campaigns, obviously, have much more sophisticated data operations than this; this is just designed to help you be a better election spectator, not run a comprehensive GOTV campaign.
Also, in some counties, the percentage ratio may change quite a bit from 2012, based on local demographics. Pennsylvania is a case in point: Montgomery County, in Philadelphia’s affluent suburbs, is diverse-enough and college educated-enough that it’s expected to give Hillary Clinton a bigger share of the vote than Barack Obama’s 57/42 margin. Luzerne County, in the state’s northeast, is white-enough and non-college-educated-enough that it’s expected to give Donald Trump a bigger share of the vote than Mitt Romney’s 52/47 loss. (If that’s worrisome, though, remember that Montgomery County is 7 percent of Pennsylvania’s total, while Luzerne County is only 2 percent of the statewide total.)
So, in each table, the first number is the county’s percentage of the statewide total. The middle column is the target Democrats should be shooting for this year, in order to eke out the barest majority at 50 percent. The right column is the 2012 results, so you can understand the baseline for my 2016 estimates. With that in mind, let’s do this thing!
ARIZONA
| County |
% of ‘12 statewide vote |
what we need to break 50% |
2012 Pres. |
| statewide |
-- |
50/47 |
44/53 |
| Maricopa |
59.9 |
50/48 |
44/54 |
| PIMa |
16.6 |
59/40 |
53/46 |
| pinAL |
4.7 |
47/51 |
41/57 |
| YAVAPAI |
4.4 |
40/58 |
34/64 |
| MOHAVE |
3.0 |
34/64 |
28/70 |
| COCONINO |
2.3 |
62/35 |
56/41 |
| COCHISE |
2.1 |
44/54 |
38/60 |
So, in Arizona, Obama lost 44 to 53. That needs to shift to 50 for Clinton to win, which would necessitate pushing Trump down to 47. (Bear in mind in reality she might need less than 50 to win, because there will be a larger 3rd-party vote this year than in 2012. I don’t have a ballpark estimate on how big the Johnson + Stein vote will be state-to-state, so, again, please treat this as a very rough projection.) Arizona is unusual among states in that most of the population is concentrated in only one county (Maricopa, where Phoenix and its suburbs are), so that makes the math somewhat easier here.
COLORADO
| county |
% of ‘12 statewide vote |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/47 |
51/46 |
| jefferson |
12.1 |
50/47 |
51/46 |
| denver |
11.7 |
73/25 |
74/24 |
| el Paso |
11.3 |
38/60 |
39/59 |
| arapahoe |
11.1 |
53/45 |
54/44 |
| larimer |
7.0 |
50/47 |
51/46 |
| boulder |
7.0 |
69/29 |
70/28 |
| adams |
6.9 |
56/41 |
57/40 |
| douglas |
6.5 |
35/63 |
36/62 |
| weld |
4.5 |
41/56 |
42/55 |
| pueblo |
3.0 |
55/43 |
56/42 |
| mesa |
2.8 |
32/66 |
33/65 |
Because Obama won Colorado, we’re adjusting downward (unlike Arizona, where we adjusted upward). Exurban Douglas County, one of the most affluent and highly educated counties in the entire country, may be the most interesting one to watch here; while it seems unlikely that Clinton actually wins here, the size of her gain over 2012 may reflect the strength of the “Never Trump” movement.
FLORIDA
| county |
% of ‘12 statewide vote |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/49 |
50/49 |
| miami-dade |
10.4 |
62/38 |
62/38 |
| broward |
8.9 |
67/32 |
67/32 |
| palm beach |
7.1 |
58/41 |
58/41 |
| hillsborough |
6.4 |
53/46 |
53/46 |
| orange |
5.5 |
59/40 |
59/40 |
| pinellas |
5.4 |
52/46 |
52/46 |
| duval |
4.9 |
48/51 |
48/51 |
| brevard |
3.4 |
43/56 |
43/56 |
| lee |
3.2 |
41/58 |
41/58 |
| polk |
2.9 |
46/53 |
46/53 |
| volusia |
2.8 |
49/50 |
49/50 |
| pasco |
2.5 |
46/52 |
46/52 |
| seminole |
2.5 |
46/53 |
46/53 |
| sarasota |
2.4 |
46/53 |
46/53 |
Barack Obama already got 50 percent in Florida in 2012 (49.9, if you want to get technical), so we don’t need any adjustments at all. Don’t expect the exact same playbook, though: it looks like Miami-Dade County (which is Hispanic-majority) may be more than 10.4 percent of the total, if its early vote numbers are any indication: its early vote numbers (750k) are only a small way behind the 2012 total vote (850k).
GEORGIA
| county |
% of ‘12 statewide vote |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
45/53 |
| fulton |
10.2 |
69/30 |
64/35 |
| cobb |
8.0 |
48/50 |
43/55 |
| de kalb |
7.9 |
83/16 |
78/21 |
| gwinnett |
7.6 |
50/49 |
45/54 |
| chatham |
2.8 |
60/38 |
55/43 |
| cherokee |
2.5 |
25/73 |
20/78 |
| clayton |
2.5 |
90/10 |
85/15 |
| henry |
2.3 |
53/46 |
48/51 |
| forsyth |
2.1 |
23/76 |
18/81 |
| richmond |
2.0 |
72/28 |
67/33 |
Georgia is a state where you might want to watch how much Clinton improves on Obama’s numbers in affluent “Never Trump” suburban counties (like Cobb) and rapidly diversifying suburbs (like Gwinnett), more so than the majority-black counties like De Kalb and Clayton, where it’ll be hard to improve on Obama’s already stratospheric performance from 2012.
ILLINOIS
| county |
% of ‘12 statewide vote |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/49 |
58/41 |
| cook |
38.4 |
66/33 |
74/25 |
| du page |
7.6 |
42/57 |
50/49 |
| lake |
5.5 |
45/53 |
53/45 |
| will |
5.3 |
44/54 |
52/46 |
| kane |
3.5 |
42/57 |
50/49 |
| mchenry |
2.6 |
37/61 |
45/53 |
| madison |
2.3 |
40/57 |
48/49 |
| st. clair |
2.3 |
48/50 |
56/42 |
| winnebago |
2.3 |
44/54 |
52/46 |
Obviously I’m not worried about Illinois from the perspective of the presidential race. I’m not really even worried about it on the Senate side; Tammy Duckworth has a pretty commanding lead. However, this is a state where the rural counties always report first, and Cook County (where Chicago is) comes last. As you’ll notice, Cook is the only county Duckworth needs to win (though she needs a crushing margin there).
INDIANA
| county |
% of ‘12 statewide vote |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
44/54 |
| marion |
13.7 |
66/32 |
60/38 |
| lake |
7.7 |
71/28 |
65/34 |
| allen |
5.6 |
47/52 |
41/58 |
| hamilton |
5.2 |
38/60 |
32/66 |
| st. joseph |
4.2 |
57/41 |
51/47 |
| porter |
2.8 |
57/41 |
51/47 |
| vanderburgh |
2.8 |
50/48 |
44/54 |
| elkhart |
2.6 |
42/56 |
36/62 |
| hendricks |
2.5 |
38/60 |
32/66 |
| johnson |
2.2 |
36/62 |
30/68 |
| monroe |
2.2 |
64/33 |
58/39 |
| tippecanoe |
2.2 |
53/45 |
47/51 |
I have zero expectation of a Clinton win in the presidential race here, but there are pivotal Senate and gubernatorial races here, with Evan Bayh and John Gregg having better-than-average shots at picking up those open seats. At the presidential level, though, the most interesting county may be Hamilton County, the affluent suburbs north of Indianapolis which are prototype “Never Trump.” For the downballot races, though, you’ll want to watch the blue-collar counties like Allen (Ft. Wayne) and Elkhart to see how far Bayh and Gregg are managing to run ahead of Clinton.
IOWA
| county |
% of ‘12 statewide vote |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
52/46 |
| polk |
14.5 |
54/44 |
56/42 |
| linn |
7.5 |
56/42 |
58/40 |
| scott |
5.7 |
54/44 |
56/42 |
| johnson |
4.8 |
65/33 |
67/31 |
| black hawk |
4.2 |
57/41 |
59/39 |
| dubuque |
3.2 |
55/44 |
57/42 |
| story |
3.0 |
54/44 |
56/42 |
| woodbury |
2.8 |
48/51 |
50/49 |
| pottawatomie |
2.7 |
44/54 |
46/52 |
| dallas |
2.4 |
41/57 |
43/55 |
Iowa is the lone state that Obama won that Clinton is likely to cede to Trump (the explanation here is that, like Ohio, it has a larger share of non-college white voters than other swing states). The Dem strength in Iowa is concentrated in its various small cities, so Clinton needs to win basically all of those counties to counteract what happens in the rural counties that you don’t see here.
MICHIGAN
| county |
% of statewide vote ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/49 |
54/45 |
| wayne |
17.3 |
69/30 |
73/26 |
| oakland |
13.8 |
49/49 |
53/45 |
| macomb |
8.5 |
47/51 |
51/47 |
| kent |
6.2 |
42/57 |
46/53 |
| genesee |
4.3 |
59/39 |
63/35 |
| washtenaw |
3.8 |
63/35 |
67/31 |
| ottawa |
2.8 |
28/71 |
32/67 |
| ingham |
2.7 |
59/39 |
63/35 |
| kalamazoo |
2.6 |
52/47 |
56/43 |
| livingston |
2.1 |
34/65 |
38/61 |
| saginaw |
2.1 |
51/47 |
55/43 |
I don’t see much polling evidence for the proposition that Michigan has gotten tight at the end, but if it does work out that way, Clinton’s main concern is mostly just making sure that turnout in Wayne County (where Detroit is) doesn’t fall any lower than the previous 17 percent baseline. Beyond that, she needs keep her percentage up in upper-middle-class Oakland County while not letting it drop too much in blue-collar Macomb County.
MISSOURI
| county |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| Statewide |
-- |
50/48 |
44/54 |
| st. louis co. |
19.2 |
62/37 |
56/43 |
| jackson |
11.3 |
65/33 |
59/39 |
| st. charles |
6.7 |
45/54 |
39/60 |
| st. louis city |
5.2 |
89/10 |
83/16 |
| greene |
4.6 |
43/55 |
37/61 |
| clay |
3.8 |
51/47 |
45/53 |
| jefferson |
3.5 |
49/49 |
43/55 |
| boone |
2.9 |
56/41 |
50/47 |
Like Indiana, Missouri isn’t going to be a presidential-level pickup unless things go totally bonkers (in a good way). But it’s the site of pivotal Senate and gubernatorial races, with Jason Kander in a tight race with Roy Blunt and Chris Koster hoping to narrowly hold the open governor’s chair. It’s a steep climb for them, especially since St. Louis city (which reports separately from suburban St. Louis County) is pretty well maxed-out percentage-wise. The best indicator of how well Kander and Koster are running ahead of Clinton may be blue-collar suburbs (like Clay County, near Kansas City, and Jefferson County, near St. Louis).
NEVADA
| county |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
52/46 |
| clark |
68.1 |
54/44 |
56/42 |
| washoe |
18.5 |
49/49 |
51/47 |
| douglas |
2.6 |
34/64 |
36/62 |
| carson city |
2.3 |
42/55 |
44/53 |
| lyon |
2.1 |
32/65 |
34/63 |
More so than any other state, the cake may already be baked in Nevada, since much of the state votes early and we know the rate at which the ballots were Dem or GOP. As you can see from the chart, Clark County (Las Vegas) and Washoe County (Reno) are nearly the whole ballgame in this otherwise-empty state, and it looks like the Democrats managed to meet the 2012 early voting rates in those two counties.
NEW HAMPSHIRE
| town |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
52/46 |
| manchester |
6.5 |
52/46 |
54/44 |
| nashua |
5.7 |
54/44 |
56/42 |
| concord |
3.0 |
63/36 |
65/34 |
| derry |
2.2 |
45/54 |
47/52 |
| dover |
2.2 |
58/40 |
60/38 |
| rochester |
2.0 |
50/49 |
52/47 |
| salem |
2.0 |
40/59 |
42/57 |
| merrimack |
2.0 |
44/54 |
46/52 |
New Hampshire, like the rest of the New England states, reports at the town level, not the county level, so don’t go looking for “Rockingham” or “Hillsborough” anywhere. New Hampshire may be either tied or a Clinton blowout depending on which terrible pollster you believe, but its largest cities, Manchester and Nashua, aren’t that much bluer than the state average, so they’re good bellwethers.
NORTH CAROLINA
| county |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
48/50 |
| wake |
10.8 |
57/42 |
55/44 |
| mecklenburg |
10.0 |
63/36 |
61/38 |
| guilford |
5.6 |
60/39 |
58/41 |
| forsyth |
3.9 |
55/44 |
53/46 |
| durham |
3.3 |
78/21 |
76/23 |
| buncombe |
2.8 |
57/41 |
55/43 |
| cumberland |
2.8 |
61/38 |
59/40 |
| new hanover |
2.3 |
49/50 |
47/52 |
| union |
2.1 |
36/64 |
34/66 |
| gaston |
2.0 |
39/60 |
37/62 |
Here’s the biggie: one of the closest states in the presidential race, plus one with pivotal Senate and gubernatorial races. (Deborah Ross is neck-and-neck with incumbent Richard Burr, while Roy Cooper looks somewhat likely to defeat incumbent Pat McCrory.) On the one hand, Clinton might actually widely beat Obama’s percentages in the biggest counties, Wake (Raleigh) and Mecklenburg (Charlotte), which are rapidly becoming more educated and more diverse. On the other hand, she could do well in all the counties above and still narrowly lose because of what happens in the smaller rural counties that don’t make this list (either the black-majority counties in the east where turnout is low, perhaps because of disruption from hurricane-related flooding, or the mostly white counties in the west which are very Trumpish).
OHIO
| county |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/49 |
51/48 |
| cuyahoga |
11.5 |
68/31 |
69/30 |
| franklin |
10.2 |
60/39 |
61/38 |
| hamilton |
7.5 |
52/47 |
53/46 |
| summit |
4.8 |
56/42 |
57/41 |
| montgomery |
4.8 |
51/48 |
52/47 |
| lucas |
3.8 |
64/34 |
65/33 |
| stark |
3.3 |
48/50 |
49/49 |
| butler |
3.0 |
36/63 |
37/62 |
| lorain |
2.6 |
56/42 |
57/41 |
| mahoning |
2.2 |
63/36 |
64/35 |
Ohio was close in 2012, and it’ll be close again this year, so the adjustment is minimal. The difference between now and then might be more about turnout rates rather than percentage; you might see Franklin (Columbus) be a larger percentage of the statewide vote than Cuyahoga (Cleveland) this year, based on which city is growing fast and which one is shrinking. One county that pundits seem to think will take a very Trumpish direction is Mahoning County (Youngstown) so keep an eye on that one (though I think Mahoning has too many non-white voters to take a big turn; the big drop will probably be smaller, whiter counties around it in the Appalachian part of Ohio).
PENNSYLVANIA
| county |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
52/47 |
| philadelphia |
12.0 |
83/16 |
85/14 |
| allegheny |
10.8 |
55/44 |
57/42 |
| montgomery |
7.2 |
55/44 |
57/42 |
| bucks |
5.6 |
48/51 |
50/49 |
| delaware |
5.0 |
58/41 |
60/39 |
| chester |
4.4 |
47/51 |
49/49 |
| lancaster |
3.9 |
38/61 |
40/59 |
| york |
3.3 |
37/62 |
39/60 |
| berks |
3.0 |
47/52 |
49/50 |
| westmoreland |
2.9 |
36/63 |
38/61 |
| lehigh |
2.6 |
51/47 |
53/45 |
| northampton |
2.3 |
50/49 |
52/47 |
| luzerne |
2.2 |
50/49 |
52/47 |
| dauphin |
2.2 |
50/48 |
52/46 |
| erie |
2.1 |
55/43 |
57/41 |
Pennsylvania’s the state that everyone always hyperventilates about and then the Democrats always win (at least at the presidential level). The Senate race may be higher drama than the presidential race here, though it looks like Katie McGinty is running almost as strong as Clinton with little coattail disparity. As I said earlier, the counties to watch here swinging sharply in the Democrat direction should be the affluent Philly-area suburbs (Montgomery, Bucks, Delaware, and Chester); the question is whether the dropoff in blue-collar places like Luzerne and Westmoreland can exert enough counter-pressure.
TEXAS
| county |
% of statewide vote in '12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
41/57 |
| harris |
14.9 |
58/40 |
49/49 |
| dallas |
8.9 |
66/33 |
57/42 |
| tarrant |
7.6 |
50/48 |
41/57 |
| bexar |
6.4 |
61/38 |
52/47 |
| travis |
4.8 |
69/27 |
60/36 |
| collin |
3.8 |
42/56 |
33/65 |
| denton |
3.0 |
42/56 |
33/65 |
| ft. bend |
2.7 |
55/44 |
46/53 |
| montgomery |
2.2 |
28/71 |
19/80 |
| el paso |
2.2 |
75/24 |
66/33 |
| williamson |
2.0 |
47/50 |
38/59 |
I don’t mean to create false hope about winning Texas by listing it here, but think of it as a dress rehearsal for 2020, when it becomes a real swing state, and 2024, when we (hopefully) win it. It’s a steep climb this year; there was a poll of only Harris County last month that was considered a bombshell because it gave Clinton a 10-point lead, but note here that she actually needs a 20-point lead to be on track to win statewide. Watch El Paso and Bexar (mostly Hispanic counties) to see if their turnout increases relative to the state, and watch Collin, Denton, and especially Fort Bend, which have a lot of Democratic upside for better percentages for Clinton; they’re much redder than they should be, based on how well-educated and diverse they are.
VIRGINIA
| county |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/48 |
51/47 |
| fairfax |
13.7 |
59/40 |
60/39 |
| virginia beach |
5.1 |
47/51 |
48/50 |
| prince william |
4.7 |
56/42 |
57/41 |
| chesterfield |
4.4 |
44/54 |
45/53 |
| henrico |
4.2 |
54/44 |
55/43 |
| loudoun |
4.2 |
51/48 |
52/47 |
| arlington |
3.1 |
68/30 |
69/29 |
| chesapeake |
2.9 |
49/50 |
50/49 |
| richmond |
2.5 |
77/22 |
78/21 |
| norfolk |
2.3 |
71/28 |
72/27 |
| newport news |
2.1 |
63/35 |
64/34 |
Virginia seems to have barely paused at “swing state” in its rapid shift from red state to blue state, thanks to accelerating Democratic strength in affluent, diverse northern Virginia (Fairfax, Prince William, Loudoun, and Arlington Counties). Nevertheless, this is one state where you’re going to see a lot of alarming votes from the Appalachian part of the state first, so it pays to know what you’re looking at. Also confusing: Virginia has both counties and dozens of independent cities, so for purposes of this chart, make sure you’re looking at Fairfax County and Richmond city.
WISCONSIN
| county |
% of statewide vote in ‘12 |
what we need to break 50% |
2012 pres. |
| statewide |
-- |
50/49 |
53/46 |
| milwaukee |
16.0 |
64/34 |
67/31 |
| dane |
9.9 |
68/30 |
71/27 |
| waukesha |
7.9 |
29/70 |
32/67 |
| brown |
4.2 |
45/53 |
48/50 |
| racine |
3.4 |
48/51 |
51/48 |
| outagamie |
3.1 |
45/53 |
48/50 |
| winnebago |
2.9 |
48/50 |
51/47 |
| kenosha |
2.6 |
52/45 |
55/43 |
| rock |
2.6 |
58/41 |
61/38 |
| washington |
2.6 |
26/73 |
29/70 |
| marathon |
2.3 |
43/55 |
46/52 |
| la crosse |
2.1 |
55/44 |
58/41 |
| sheboygan |
2.0 |
42/57 |
45/54 |
| -- |
-- |
-- |
-- |
Finally, let’s look at Wisconsin. This isn’t so much a concern at the presidential level as at the Senate level, where Russ Feingold’s once-likely pickup has gotten much tighter in the closing weeks. Also, it just wouldn’t be right to do a project like this without making a few wisecracks about Crucial Waukesha County. At the presidential level, that may be the most interesting county, because it’s pretty much the prototype for “Never Trump:” a well-educated, dark-red county where Trump is particularly distasteful (this was Ted Cruz’s stronghold during the Wisconsin primary). In the Senate race, though, the main concern may be making sure that Milwaukee and Dane (Madison, where Univ. of Wisconsin is) Counties turn out at a high enough rate and that they don’t lose their large share of the state’s votes.
