The Recount: There are only two possible outcomes: 1) confirmation of the political position of a plurality of Wisconsin voters or 2) election fraud on a scale not previously seen, at least not anytime recent which would be in the past hundred years or so. The latter outcome would shake the entire US electoral system. The former should shake the Democratic Party to wake up and smell the coffee.
My guesstimate of the odds that it’s #2 is much lower than the pre-election odds for a Trump win. The numbers simply don’t support a plot to fix the vote in WI because it would have to involved all but four counties. And two of those four are solid Republican counties. Let’s walk through the numbers.
2012:
Obama: 1,613,915
Romney: 1,408,746
Total: 3,022,696 votes cast for the two major political party’s nominees.
Winning vote margin: 205,204
2016:
Clinton: 1,382,120
Trump: 1,410,458
Total: 2,702,578
Winning vote margin: 28,338
Voter participation on the presidential ballot line in 2016 was down by 110,310 compared to that of 2012. That’s a little more than half of Obama’s 2012 vote margin of 205,204. What can’t be known is the actual number of 2012 voters that passed on 2016 and who they voted for in 2012. What is known is that the drop-off/drop-out occurred in almost all WI counties, regardless of which candidate won the county in 2012.
Could that net drop-off (newly eligible voters less newly dead previous voters) reflect the impact of Wisconsin’s new voter ID requirements? Sure. However, as the drop-off/drop-out voters in 2016 were not limited to states with voter suppression efforts that case is weak at best. In addition, GOP voter suppression targets the poor and POC most likely to vote Democratic and therefore, the impact would have to be seen most dramatically in communities with high numbers of poor, POC, and Democratic voters. No evidence of that in the WI numbers.
For example: Waukesha county. Suburban, white, not poor, and solidly Republican. Election results:
2012:
Obama: 77,617
Romney: 161,567
other: 1,900
total votes: 241,084
2016:
Clinton: 79,199
Trump: 145,519
other: 11,551
total votes: 236,269
A decline in total voters and Clinton received more votes than Obama in 2012 and Trump received fewer votes than Romney. Also note that the decline for Trump was larger than the increased number of third party voters. Thus, Clinton had to have gotten some Romney voters in numbers greater than a loss to GRN voters. This was precisely the strategy of Clinton’s campaign. Peel off middle-class, white, suburban Romney voters by highlighting the degree to which Trump is disgusting. Two problems with that strategy: 1) voters in this demographic are as reliably GOP voters as AAs are Dem voters. 2) secondarily, voters in such communities have a high vested interest in maintaining the status quo over “change.” Had Clinton pushed harder on being the status quo candidate, she might have gotten more Waukesha county votes, but for every vote gain there, she would have lost an equal or larger number of “change” voters elsewhere.
Another example: Dane County. Less white, less wealthy, high public employment (IOW more job security), and reliably Democratic. (An obvious target for the Scott Walker administration.)
2012:
Obama: 215,389
Romney: 83,459
other: 3,997
total votes: 302,845
2016:
Clinton: 217,506
Trump: 71,270
other: 15,953
total votes: 304,729
So, in two out of the three largest counties in WI, Clinton did better than Obama. How about the largest county, Milwaukee?
2012:
Obama: 328,090
Romney: 158,430
other: 4,424
total votes: 490,944
2016:
Clinton: 288,986
Trump: 126,901
other: 19,893
total votes: 434,970
Aha!? Uh, no. Trump’s net gain in Milwaukee County over Mitt’s performance was only 6,765 votes, and recall he needed a statewide gain of 205,204 plus one and Dane and Waukesha put him another 28,237 votes behind Mitt. The number of votes in all but eleven counties was down from that of 2012. The pattern seen in most counties is similar to that of tiny Florence county:
2012:
Obama: 952
Romney: 1,643
other: 20
total votes: 2,615
2016:
Clinton: 666
Trump: 1897
other: 88
total votes: 2,651
Didn’t matter if the total number of votes increased or declined or the county went Democratic or Republican in 2012. In the aggregate, Trump held the Romney’s vote totals and Clinton didn’t hold Obama’s vote totals.
Interestingly, in those counties where Trump had gains over Romney’s number, Feingold posted better numbers than Clinton. Alas, not as good as Obama’s ’12 numbers. However, it could have been good enough. In those counties where Feingold received more votes than Clinton (all but seventeen counties), we see:
Net drop-off/opt-out on Senate ballot line = 1,619
====
Trump: 666,544
Johnson: 667,122
===
Clinton: 671,479
Stein: 16,334
(total Clinton and Stein = 687,813)
Feingold: 699,909
===
LIB & crank on POTUS line: 59,246
LIB on Senate line: 45,137
(total Trump and net LIB & crank = 680,642)
Johnson: 667,122
Thus, Feingold had to have picked up some of those Trump, LIB, and/or crank voters in those counties. With regard to the counties where Clinton did better than Feingold (The Seventeen), the pattern was different:
Net drop-off/opt-out on Senate ballot line = -7,706 (an increase for senate line voters)
====
Trump: 743,914
Johnson: 812,141
===
Clinton: 710,541
Stein: 14,646
(total Clinton and Stein = 725,287)
Feingold: 680,597
===
LIB & crank on POTUS line: 63,397
LIB on Senate line: 42,154
(total Trump and Net gain LIB & Crank = 746,530)
Clearly, Johnson picked up Clinton voters in these counties. In these counties, had Johnson been limited to the Trump plus the net LIB and crank votes and had Feingold held the Clinton and Stein votes, Feingold would have won.
If you’re wondering how the 2016 senate vote in these counties compared with the 2010 senate vote, the answer is about the same. Feingold carried the aggregate of the same counties in 2010 and 2016. Clinton also won by 4,935 votes in Feingold Country.
So, how did Obama fare with the same split in the WI counties?
2012 Feingold Country:
Obama: 795.405
Romney: 617,989
Obama vote Margin: 177,416
===
2012 The Seventeen:
Obama: 818,545
Romney: 790,757
Obama vote margin: 27,788
====
2016 Feingold Country
Clinton vote margin: 4,935 (total votes shed 172,481)
2016 The Seventeen:
Trump vote margin: 33,273 (total votes shed by Clinton 61,061)
I can wrap my head around why there were a couple hundred thousand Obama ’12 voters that went with Trump in ’16. I can also wrap my head around Stein voters choosing Feingold. Also the twelve thousand in Feingold Country that went with Trump, LIB, or crank and Feingold for Senate. What doesn’t compute for me are the (roughly) forty thousand plus Clinton voters in “The Seventeen” that chose Johnson over Feingold. Were they “divided government” fetishists? IOW, if Clinton were to win (as everyone expected), they preferred a WI GOP Senator to keep her in check. If Trump had been the expected winner, would they have preferred a WI Democratic Senator to keep Trump in check? Voter overthink? Or dare I suggest that there was some vote theft from Feingold to Johnson in The Seventeen. (Isolated to a few counties and at variance with the pattern of the other counties. This is how a vote theft operation would appear in the results. Not that such a deviation is evidence of vote theft or is unnatural.)
What this little exercise reveals to me is that Feingold’s pre-election poll numbers, consistently near or over 50% were probably accurate as to voter intent. Then as they soured on Clinton, they took it out on Feingold. Hillary, the anti-coattail wonder.
Note: Most of numbers were pulled from the Wisconsin Elections Commission website Those pulled from other sources differed by a few votes but not enough for me to redo my spreadsheets.
It’s the Water. Or the Wealth. Or something something.
The Seventeen may qualify as “Walker Country,” but I haven’t run any numbers of that; so, will stick with The Seventeen. Fifteen are the along the eastern side of WI, either on the water or contiguous to a county on the water, and run from north to south. One is tiny and north central and one is western central. The range in median income from #1 to #72 (last place).
Income rank in WI
#1 – Ozaukee
#2 – Waukesha
#5 – Washington
#6 – Door
#8 – Calumet
#11- Outagamie
#12- Brown
#13- Walworth
========
#15- Winnebago
#16- Racine
#17- Pierce
#18- Kenosha
#21- Fond Du Lac
#22- Manitowac
#24- Sheboygan
#40- Milwaukee
#72- Menominee
The Wisconsin statewide median income is at the break above Winnebago. So, The Seventeen do skew towards the upper and middle income range.
As Clinton only carried one county in the 2016 primary, Milwaukee, there’s no correlation from that to disfavoring Feingold in the Senate race. And overall, the counties where she did better in the primary don’t predict The Seventeen either.
Of The Seventeen in the general election, Clinton carried Menominee and Milwaukee.
Interesting WaPo analysis: Trump won in counties that lost jobs to China and Mexico. (Uses actual data and not the idle and baseless accusations that went into its blacklist of alternative media sites.)
So, “China is taking your jobs” beat “Putin/Russia is hacking US computers.”
Would have been fun to see Sanders going after Trump on his anti-China/jobs rant considering that Trump took advantage of low manufacturing costs for his own products.
Really well done.
Do you have your WI county data in a spreadsheet?
I am comparing county level election result data against change in immigration for Iowa. Next up is Mn, Mi, WI and FL.
Right now my sheet is census data by race and education by county in Iowa, matched against ’12, ’08 and ’04 county level results.
I do. And haven’t a clue how to post a spreadsheet here.
The WI Election Commission is posting the results of the recount in an on-line (read only I think) spreadsheet by precinct. That’s the sort of election results transparency that should be done everywhere.
I can’t stress enough that my sort was based on a single variable: counties where Feingold received more votes than Clinton. And that came after I had observed that in some of those counties the sum of Clinton and Stein votes was less than Feingold’s. It only became interesting when I did the same county split for 2010 and 2012. Surprised me that the total votes in the two piles were roughly similar and that most of Obama’s winning vote margin came from those in the “Feingold country” pile.
Other than the three large counties that I included in the diary, I haven’t looked at the location or demographics of any of the other counties because my focus was on candidates Clinton and Feingold. And I was also trying to wrap my brain around why the polls consistently had Feingold with a solid lead (50%) and then he fell way short of that percentage.
Here is a place. I have some of my Iowa data there and will be building it out:
https://docs.google.com/spreadsheets/d/1EytIucVwhu898DF9zBlE3cLAZVvuq1g5WdNrr65vC3g/edit#gid=0
Matt Taibbi in Rolling Stone (that along with 99% of newspapers and periodicals endorsed HER over Bernie bc she could win and yet lost to Donald — freaking — Trump) — Bernie Sanders: Where We Go From Here.
Bernie still talking sense and not the delusional politico-babble that most Democratic politicians and their sycophants spew.
One very clever and opaque dig at the Clintons by Sanders during the interview. Otherwise, he spoke of the reality that exists today.
Disagree that Democratic politicians are timid. They’re just not as adept at serving their masters and conning their voter base as the GOP is.
Yeah, the Clinton News Network (aka CNN) really showed how to elect a president. But they did a bang up job on getting HER nominated.
It makes sense to me, and is what I had expected. They are probably Republicans strongly opposed to somebody of Trump’s temperament. Also, a small number of these might have been resistant to Trump since he bashed whoever they voted for in the primary – were these Kasich counties by any chance?
Also, were those seventeen counties mostly suburban? Hillary’s base is very different from Feingold’s, which probably hurt him a lot. States where Hillary had coattails (or more likely, Trump had negative coattails) were Hispanic.
They are probably Republicans strongly opposed to somebody of Trump’s temperament.
I considered that and ended up rejecting it because it would mean crossing party lines for a Democrat that wasn’t expected to need their vote and they had other options available to indicate their disgust with Trump, including leaving the top line blank. Now if Johnson were on the left side of the GOP, a Clinton-Johnson voter would be consistent. But he’s pretty far to the right in the GOP.
Ah, you noticed that I didn’t name The Seventeen that displayed a single difference from the other WI counties — specifically more votes for Clinton than Feingold. (Also confirmed in the aggregate that went with Feingold in 2010.) They range in size from small to large. Obama carried seven of them in ’12 and did very well in three others. Clinton carried two of them. Democrats do struggle with white suburban voters, but Clinton didn’t break through that wall to any measurable degree and lost big time in the hinterlands. Maybe those Clinton-Johnson voters are anti-Semitic.
Could you list those seventeen counties? A map would be helpful, but based on other maps I’ve seen, a lot of them likely are in southwestern Wisconsin and are in the same area as a lot of Obama-to-Trump flips in Iowa, Minnesota, and Illinois.
If there are Romney-to-Clinton flips, they likely are in the first ring of Milwaukee suburbs; my guess is that there aren’t any.
The voting patterns of one county are not independent of those of contiguous or nearby counties.
Why? I’m very weary of the demographic games both parties play. It’s a distraction from the meaning of politics and politicians to the lives of real people and exacerbates the political polarization that has increased over the past few decades.
What were voters in all but seventeen WI counties responding to that in the aggregate caused them to choose Feingold over Johnson? Re-confirming their 2010 decision. His honesty, decency, and competence? Yet, not as valued in The Seventeen. It was also where Obama performed best in the state in 2012 and where Clinton collapsed the most. Seems to me that given a choice between dishonesty, indecency, and incompetence and faux honesty, indecency, and incompetence that they’ll go with the latter.
Johnson outperformed in western and northern Wisconsin. If you compare his numbers to Baldwin in 2012 it would appear that independents who preferred Baldwin in some of those counties chose Johnson this time around. I think it probably made the difference.
Both elections favored the political newcomer, more-so in Johnson’s case, to the experienced politician. Feingold couldn’t credibly run as an outsider on the same ticket as HRC. It didn’t help that people are attracted to Trump’s “drain the swamp” message.
Baldwin did well where Obama did well. A few votes shy of Obama’s totals but close enough that there was no need to present her Senate race as unique from Obama’s ’12 run. In one county did she perform better than Obama by two votes; otherwise she ran a few hundred or thousand votes behind Obama. The bulk of both of their winning voter margins were generated in what I’m referring to as “Feingold Country.”
Yep.. I think Feingold was unlucky in that he got caught in a change election on the same ticket as HRC.
What you refer to as “Feingold country” is now where Democrats are weak in Wisconsin. Republicans here have been quite effective in framing state politics as the rest of the state against Madison and Milwaukee.
However, I don’t think that’s really a reflection on these particular independent voters I mentioned, who preferred Baldwin over Tommy Thompson.
No, you misunderstand. Feingold won in “Feingold country” in both 2010 and 2016. It was Hillary that collapsed there.
It’s The Seventeen (roughly half the WI population) where Democrats are weak.
I think your data about areas where Feingold performed well in 2010 and 2016 and when compared to Hillary in 2016 is sound. Hillary definitely collapsed there. Not something I take issue with. She lost my county to Bernie by ~15 percent.
Seeing which counties they are might allow other factors that are common to them be examined. Messaging involves more than the sender and the message; it involves the medium and the recipient. What is the reality of the lives of of people in these 17 counties and not the others in Wisconsin that could account for the voting pattern you describe?
Yes — and I’ll post the list shortly — but I wanted to first pose a non-demographic question. Approximately half of WI voters living in all but seventeen counties preferred Feingold to Johnson. The sum of the votes from these counties also gave Obama must of his WI vote margin in ’12 and Feingold carried this bloc of counties in ’10. It’s also where Clinton collapsed. So, the meme that it was rural white voters that defeated Clinton isn’t true. By a tiny number of votes, she too carried this bloc.
The best analysis thus far says ITS EDUCATION and not location that is behind the shifts.
But there is work to do and not many have gotten to the county level data.
Of course after that we get to the precincts where the real information resides.
This.
There is an area long the intersection of Iowa, WI and MN that saw significant shifts.
I spent the morning looking that the Iowa County results which are just absolutely shocking. There isn’t any real reason why Minnesota, Wisconsin or Michigan would be all that different. The only difference is the ration of rural to urban.
There is no other way to put this: the Democratic Party is on the verge of a complete collapse in the rural Midwest.
“These issues apply to more than rural and semi-rural areas. Trump’s message – that firms need to consider something more than bottom line – resonates in middle and upper-middle class households as well. They know that their grip on their economic life is tenuous, that they are the future “low-skilled” workers. And they know they will be thrown under the bus for the greater good just like “low-skilled” workers before them.”
http://economistsview.typepad.com/economistsview/2016/12/desperately-searching-for-a-new-strategy.ht
ml
Marie3, I think your discussion of demographics and voting trends is pretty interesting, but I don’t see how it bears on the issue of possible fraud via, say, tampering with voting machines.
When Steven D wrote a couple of diaries in the spring alleging a widespread conspiracy to steal votes from Sanders on Hillary Clinton’s, he was supporting a line of argument grounded in statistics and claims about the, uh, Biblical inerrancy of exit polls. Some readers here, including me, strenuously disputed Steven D’s arguments. You did not. It was never quite clear to me if you actually supported the claims that Steven D was making or if you were simply playing devil’s advocate by challenging the people who disagreed with Steven D.
Now along comes a challenge to the accuracy, perhaps the legitimacy, of the general election results in three states. The challenge calls upon statistical arguments. You don’t really address those arguments but instead, for reasons I don’t follow, get into some interesting but perhaps irrelevant discussion of differences between 2012 and 2016.
The Green Party updates showing up in my e-mailbox state that there is evidence developing of hacked voting machines in Wisconsin, at least. A fundraising tactic by the Greens? Or something real?
2nd paragraph should read “on Hillary Clinton’s behalf”.
I’m going to treat your inquiry as being authentic, but if you then veer off into your standard get me to respond as a ploy to attack or nitpick me or start a boring and meaningless argument again, this will be the last time I ever give you the benefit of doubt.
iirc Steven D’s report was more than numbers or statistics. There was a high correlation between Sanders’ performance in precincts or counties that used voting methods with a paper trail and those that lacked a paper trail. That was interesting enough that it should raise questions about non-paper trail voting machines. Those that attacked this information, and by extension Steven D, as hogwash were nothing more than partisan hacks. So, I may have defended Steven on this point.
To jump from this data to accusations of election fraud is too big a leap for me. Not that they may be correct, but only that it’s not sufficient evidence. So, that’s why I wouldn’t have taken a position on that point. This is a failing of all of the claims of vote theft through electronic voting machines. Again, not that such vote theft can’t be done, but it’s a lot more complicated to accomplish than this sort of paranoia suggests. (It’s also why I easily dismissed the Clinton/DNC/Democrats claim that Putin was going to hack and steal the 2016 election. Unlike Steven D’s report, they didn’t even have any data or evidence that something like that was in the works or could actually be done.)
Assume motive for election theft is a given, we have to ask about means and opportunity. Counties and some cities manage our elections. For better or worse, they choose different methods and machines, if any, for voters to use. If the machines aren’t linked to a central computer on election day and they have a paper trail, the ability to rig the outcome is much more difficult and requires so much coordination among election workers in various locations that it’s not a reasonable scenario. Somewhat less cumbersome to rig if the voting machines aren’t on-line to a central computer on election day but still it’s a lot of machines that have to be rigged.
Next issue is where the votes are counted. It’s my understanding that in WI it seems to be done at the precinct or multi-precinct level and then reported up to the county administrator who then reports the results to the state election commission. That’s a lot of human hands touching the counting and results for a single plot to rig the election to operate seamlessly and undetected in more than a couple of locations. Absentee ballots that all go to a county/city central counting location is different and why it’s one of the easier points in the process to rig.
Wrt Steven D’s report, the absence of an audit trail means that there was no way to verify the questionable results. So, it wasn’t worth my time to engage in idle speculation. His case would have been stronger if it had included specifics down to the precinct level and how and where the votes were counted and reported upstream and the results compared to the voting in those locations in prior elections. That’s a huge amount of information to collect, much of it difficult or impossible to collect, and then sift through.
The alternative is to identify weak links in the chains and where motive, means, and opportunity are most prevalent. IOW, think like a crook. I once did employee theft loss control. No way to close off all opportunities for a single hit. However, there are many ways to limit the amount an employee could get in one hit. The opposite extreme is an employee regularly stealing a little bit over an extended period of time. Those are never detected except by luck or accident or by a highly perceptive accountant and then always after the theft has been on-going for an extended period of time.
So, for election rigging, does the thief steal big in a large county or steal small in all counties? The latter is difficult to manage but essentially undetectable. In WI 2016 there is no evidence of large theft in the presidential election in the large counties. The results are fully in line with those of 2012. Even Milwaukee Co with a significant drop in voter turnout only change the vote margin for Trump by 5,000 over what Mitt received. That leaves us with the alternative of theft spread throughout all of the other counties in WI.
Conspiracies are an ordinary, everyday fact of life (which is why those that use CT as a pejorative are full of crap). However, a one-shot, widely dispersed conspiracy has too many weak links to succeed. (Unlike an institutional “conspiracy” that can go on for decades, outlive the original architects, and become accepted as just the way things are done in the organization. Unless or until someone new steps in and says WTF.) So, in the WI 2016 election, I reject such a conspiracy. They would have needed to steal too a large number of votes and in too many locations to succeed. However, it wasn’t too large a number of voters to have gone off Obama and choose one of the alternatives to Clinton who they weren’t wild about to begin with.
You can follow the WI recount through the daily spreadsheet posted at the Wisconsin Election Commission by precinct or ward. I’m not compulsive enough and not being paid to monitor this. From a brief scan, it looks exactly like what is seen in a recount where there is no systematic rigging of the results. A vote or a few votes up or down for the candidates or no change.
My diary used the election result numbers to pose serious political questions. While ’16 voters in most WI counties preferred Trump to Clinton, half the voters in a group of counties preferred Feingold to Johnson and favored Feingold in larger numbers than they did Clinton. Why? On the presidential ballot line, the sum of the votes from these counties looked as if there had been a large shift among these voters from D to R. Yet on the Senate ballot line, they were midway between ’10 and ’12 and even in ’10, Feingold had carried this bloc of counties. Why would the people in these counties vote more similarly to The Seventeen on the POTUS ballot line but still hold Feingold in high regard?
The analysis Steve D was so bad – remember its origin is a bunch of Rand Paul fanatics – that I really lost respect for Caucus99 as a whole.
If you looked at the data you realized it was nonsense. There was stuff in about NH precincts which actually showed the opposite of what they are saying.
One other item. Solidly partisan counties are the most likely to fraudulently increase the count for their party. If they are rural counties, they are less likely to be discovered because the individual contribution to the total vote is less easy to track back. That is why late-reporting counties (or precincts) are always suspect and subject to recounts if the election is close.
It is what motivates the NC GOP to force a recount of Durham County in the governor’s race, for example.
Bernie’s book tour in the west — MarinIJ – Bernie Sanders in San Rafael: Democratic elite to blame for Trump victory
A reference note: Clinton stomped Bernie in Marin. It’s HER kind of county. Working class people can’t afford to live there.
Voter over-think bc everybody is a political pundit.
Another case for preference voting.
Voters don’t “over-think” in preference voting?
Maybe, but they have a fall back.
How so?