Wednesday, November 16, 2016

Mea Maxima Culpa! Election 2016 Post Mortem

Wow - I really blew this one!! There is safety in numbers as just about everybody else in the punditry world did as well. Suffice to say, there was a much greater antipathy to conventional politicians than almost anyone anticipated and 2017 will welcome in President Trump.

Reasons Trump won
  • Kept a very simple message: the economy, immigration, and corruption. He stayed on message despite numerous misstatements and less than stellar debate performances.
  • He attracted unusually large numbers of blue-collar whites in the rust-belt. This was the margin of victory as many counties in these states shifted to the GOP for the first time in 30 years.
  • Trump's own failing in both business and his personal life did not have a major effect at the end of the day.
  • Public opinion polling was incorrect almost across the board. Nationally, polling was about three points off and it was further off in much of the state polling.
  • The "shy" Trump vote. It is possible that the "Bradley Effect" may have moved the race a point or so in close states as Trump voters either unwilling to participate in surveys or intentionally misleading pollsters due to social stigma. This can't be proven, but may be a part of this story.
  • Lower voter turnout among blacks and young people as compared to 2012.
  • Trump's base supporters were very loyal - and they voted in large numbers.

Reasons Clinton lost
  • The email scandal - Clinton simply was not seen as honest or trustworthy. The FBI Director's letter 11 days prior to the election reminded people of this. The Clinton Foundation's perceived ethical issues and both Bill and Hillary Clinton's cumulative (alleged) scandals since the 1990's were brought into focus by this as well.
  • She did not pay enough attention to jobs and the economy - people are hurting in many parts of the nation and they did not perceive her addressing their concerns. When she said that things in the nation are well they vehemently disagreed.
  • The Clinton campaign thought the "blue wall" was solid. Despite some worrying reports from the field they reacted too late in Michigan and were absolutely shocked to lose Wisconsin.
  • The challenge in the primaries from Bernie Sanders made her shift positions and appear to have no core beliefs.
  • Constant distraction by the Wikileaks dumps.
  • Arrogance - pure and simple. The Clinton campaign thought this election was already sealed and they appeared to let up on the campaign toward the end. Relying on surrogates and celebrities probably won them few additional votes. In effect, they rode a victory lap on a flat tire.

Now to look at my own predictions from last week... (warning disturbingly incorrect predictions ahead!)

Clinton is going to win due to the following factors.
  • A unified Democratic Party (Bernie supporters have come home) and a far less unified Republican Party with multiple prominent Republicans either publicly supporting her candidacy or not endorsing Trump. Clinton is likely to garner the support of 90% or more of Democrats. If Trump manages only 85% of Republican voters it will give her enough of an edge in the swing states to win the majority of them.
According to the exit polls Trump actually drew 90% of GOP voters while Clinton drew 89% of Democrats. This was certainly enough to help sway close states such as PA, FL, and NC. 
  • Independents who usually lean to the GOP appear to be breaking slightly to Clinton. More bad news for Trump.
Independents broke for Trump 48-42.
  • Superior Clinton ground game in all respects. This is important in driving key supporters to the polls in both early voting and on election day. Data on early voting suggests that it is on par with 2012 which produced Obama victories in almost all of the swing states. by contrast, Trump has publicly shown his disdain of data and organization. This will cost him dearly.
Another miss here. Clinton's data driven approach failed to turn out key segments of her coalition (African-Americans) in sufficient numbers contributing to the losses in NC, PA, MI, and WI. Trump's "gut-feel" approach succeeded in driving far larger number of blue collar whites to him in the rust belt. Clinton ignored this group at her peril and it cost her dearly.
  • Clinton's policy positions (overall) are more in-line with the majority of voters. This varies of course with Trump's positions on the economy and jobs being preferred by a large number of voters - but the urban, college-educated, female, and minority coalition that form the backbone of the Clinton candidacy will outweigh that of the Trump coalition of white, male, religious, and less educated voters.
This was not inherently incorrect, but many voters were not voting on issues or by party per se...
  • Expanding on the aforementioned point, Trump is getting hammered with female voters by over 20 points who have both a chance to make history and reject Trump's perceived chauvinism. Trump leads among men, but by a far smaller margin than Romney did four years ago.
Clinton indeed did win female voters, but "only" by 14 points. Normally this would be a significant win, but Trump won men by 14 points as well so the gender gap was basically a wash. Delving a little deeper into this, Trump won white women 53-43 and white men by a massive 63-31 margin. Clinton won white college educated women by only 5 points. College educated whites voted for Trump 49-45 which was not in line with the pre-election forecast.
  • The economy, while not booming, is perceived as being good enough to keep the Democrats in power. There are large pockets where the economy is doing poorly mainly in large segments of the Midwest (Ohio, Michigan) and this benefits Trump. But it is not going to be enough to win enough electoral votes in the end.
Wrong again Prof Brad - the perceived poor economy in Middle America was the major reason Clinton lost in much of the rust-belt.
  • In an ordinary election year the base factors would tilt slightly to a GOP win after eight years of a Democratic administration. But this year Trump has alienated and has made himself unqualified as act as President to a vast segment of the electorate and suffers from very high disapproval ratings. Hillary suffers high disapproval as well, but she is seen as being qualified for the job by more voters. This is very important as it gives wavering Republicans a place to go to if they cannot bring themselves to support Trump.
Despite Trump's disapproval of 60%, many voters overlooked this. The GOP largely came home. Clinton's disapproval numbers were not too far behind at 54%.
  • As predicted, the third party candidacies of Johnson and Stein have faded and they will have little overall effect on the election.
Both Johnson and Stein underperformed. They may have swung a state or two given their vote in some of the close states, but there is no data to support that as of yet.
  • There is going to be a very powerful anti-Trump wave in the Hispanic community. It will put tossup states like Nevada and Florida in Clinton's corner, will keep states that lean Democratic like Colorado in Clinton's column and  and will diminish Trump victories in Arizona and Texas. This is a major and existential problem for the GOP going forward. If Hispanics solidify 80% or more support for the Democrats, the Republicans will be hard pressed to win another presidential election in the foreseeable future.
I still think that Hispanic voters are an existential problem for Republicans. This is demonstrated by the less than landslide Trump victories in AZ and TX. But if the exit polls are to be believed, Trump lost Hispanics 65-29 which is quite close to Hispanic breakdown margin in 2012.
  • Finally, the Democratic "blue wall" of states that give them a comfortable base in the Electoral College appears to be largely holding. This gives them a base of over 250 EVs which means that only one or two of the swing states needs to be won to ensure victory. The GOP literally needs to run the table to win.
And indeed he did pull the inside straight! The blue wall was not only breached, but broken. PA, MI, OH, and WI all flipped from Blue to Red. PA and MI have not voted GOP since 1988. Wisconsin has not voted Republican since it supported Ronald Reagan in 1984.

Electoral College Final Results 2016

Popular Vote Breakdown
Clinton wins the popular vote by 5 points, but fails to break 50%.

Clinton      49
Trump       44
Johnston    5
Stein           1
Other          1

I was right on Clinton being popular vote winner (mainly through her landslide wins in blue CA and NY). These results may adjust a bit as the final popular vote tally will not be certified for a couple more weeks.

Clinton      47.8
Trump       46.7
Johnston    3.2
Stein           1.0
Other          1.0

Lets see how I did on the individual state calls...

Arizona - Typically a GOP stalwart, Hillary has made a play due to it's growing Hispanic population. AZ is trending more purple of late, but it is probably a bridge too far - this year. Prediction: Trump by 5

I was pretty good on AZ. It did get closer for Clinton, but not close enough. Actual: Trump by 4

Florida - This is a must-win for Trump as the 29 electoral votes are critical for any chance he may have. Both candidates have poured money and resources into the states. Results from the bellwether counties of Hillsborough and Polk in central Florida will give an early indication of who will prevail. There is a large white rural demographic in northern Florida and the Panhandle that will run up the score for Trump. Clinton will clean up among the growing Puerto Rican community. Elderly voters will favor Trump, but the margins may not be enough. My sense is that this may look a lot like 2012 where Obama squeaked out a 0.9% victory. This is a knife's edge and will be decided who has the better ground operation. Thus I am calling it for Clinton. Prediction: Clinton by 2

Florida behaved like the swing state it is with the usual regions supporting the GOP or the Democrats. However, Trump overperformed in some swing regions and it was enough to swing it to him. Actual: Trump by 1.3

Georgia - Like Arizona, Georgia is trending more purple as the Atlanta metro area has a large population of college-educated whiles and there is a sizable African-American community that will swing strongly to Clinton. But also like Arizona, it is still a fundamentally Republican state with demographics that favor a Trump victory. Prediction: Trump by 5

Georgia stayed GOP as Trump ran in totals in the rural pars of the state. As expected, Clinton did well in metro Atlanta. Actual: Trump by 6

Iowa - This state voted twice for Bush by a close margin and was won by Obama twice by comfortable margins. However, the demographics favor Trump and he is leading in all of the late polling. Prediction: Trump by 4

Iowa defied expectations with Trump winning all but 5 counties. Actual: Trump by 9

Maine - Using the District plan for apportionment of Electoral Votes, Maine splits its EVs by the winner of each congressional district. The statewide winner receives 2 EVs  and then the winner in each of Main'es two districts receives one EV per district won. Clinton will win statewide in Maine easily based on a victory in relatively wealthy CD 2 in southern Maine. But Trump is striving for a slim victory in CD 1 in poor and rural northern Maine. Prediction: Clinton by 6 statewide (3 EVs)Trump by 2 in CD 2 (1 EV)

I got the split correct in Maine! Actual:Clinton by 3 statewide (3 EVs)Trump by 10 in CD 2 (1 EV)

Michigan - Michigan has become a state that both candidate are working in over the past two weeks. Clinton has ramped up operations in this usually-blue state as a struggling economy and large white and blue collar population are making possible a Trump pickup.  Unfortunately for Trump, Clinton will win a massive victory in minority Detroit and the educated white Detroit metro area which will outweigh his vote elsewhere. Prediction: Clinton by 4

Wow - Michigan did surprise.Clinton did indeed win the metro Detroit area, but the turnout was down just enough to be offset by Trumps running up the vote elsewhere and flipping many counties that Obama won in 2012. The white blue collar workforce in this struggling state made the difference. Actual: Trump by 0.3%.

Nebraska - Like Maine, Nebraska uses the District Plan to allocate EVs. Obama won the single urban congressional district around Omaha in 2008 to steal that electoral vote. Clinton is making a play for it, but I think Trump squeaks it out to carry all 5 of Nebraska's EVs. Prediction: Trump by 12 statewide; Trump by 2 in CD 2

Big Trump win in Nebraska including CD 2. Actual: Trump by 26 statewide; Trump by 3 in CD 3

Nevada - Won by Obama twice, this formerly red state is trending blue due to demographic changes. The Hispanic vote will secure a victory for Clinton as Nevada follows New Mexico away from the GOP due to a declining white population, Trump's stance on immigration certainly does not help here either. Prediction: Clinton by 4

The Democratic machine was able to deliver Nevada for Clinton, albeit by a close margin. Actual: Clinton by 2.5

New Hampshire - New Hampshire is always in play and Trump over-performed  in the February Primary. Sanders pummeled Clinton here as well in February...but my gut and the polling is indicating a modest Clinton win here. Prediction: Clinton by 4

NH was won by Clinton in a razor thin margin of 3,000 votes. Actual: Clinton by 0.3%

North Carolina - Another red state that is turned purple in recent years, the 15 EVs in North Carolina are being bitterly fought over. Like Florida, this is a close call and wont be decided until late in the evening or the next day. Clinton will barely squeak it out with high minority turnout and the winning northern Whites with college degrees who have relocated to this booming state. Prediction: Clinton by <1 span="">

North Carolina surprised and went to Trump despite all the attention by Clinton and the President in the last week of the campaign. Actual: Trump by 4

Ohio - Ohio is tailor made for Trump (blue collar, white, and smaller percentage of college grads) and usually is a must win for either party. Luckily for Clinton, she can afford to lose Ohio due to her strength elsewhere. Watch the vote totals late in the evening from Cleveland and Cincinnati to see if she can catch Trump who will build a large lead in the more rural and blue collar regions. My belief is that the cities won't be enough o put Clinton over the top. Prediction: Trump by 3

Trump dominated and flipped many hardcore Democratic counties in Appalachian Ohio. This was not that close in the end. Actual: Trump by 9

Wildcards - states that are not likely to swing, but one never knows:

Oh, was I wrong here!

Pennsylvania - Described by pundit James Carville as "Philadelphia and Pittsburgh on the ends with Alabama in the middle" Pennsylvania has long been sought after by the GOP who have not won here since 1988. Romney made a big push here in 2012 and lost by 5 points. Philly and the suburbs will be a lonely blue island in an otherwise red sea, but the suburban female vote will be so overwhelmingly pro-Clinton that she should not be too worried here. Prediction: Clinton by 6

A surprise Trump win by racking up very large margins in rural PA, taking industrial Erie County and doing just well enough in suburban Philly. Clinton did better than Obama in the suburban counties, but did not do as well as Obama in central city Philadelphia. Actual: Trump by 1.2

Virginia - Yet another solid Republican state which has gone blue due to the booming and highly educated DC suburbs. If Clinton's running mate Tim Kaine was not a sitting Virginia senator it may have been marginally in play. Prediction: Clinton by 7

Clinton managed to hold onto VA due to a strong suburban DC vote overwhelming Trumps's vote downstate. She did slightly better than Obama in 2012, thus turning VA a bluer shade of purple. Actual: Clinton by 5

Texas - What?! Yes, this GOP anchor is slowly eroding as the Hispanic population continues to grow. Trump will win, but his margin will not be close to Romney's 16 point win in 2012. Prediction: Trump by 9

As predicted, Trump did not do as well as he should have in Texas - warning sign for the future... Actual: Trump by 9

Wisconsin - Narrowly won by Gore and Kerry, Obama won here easily despite local Congressman Paul Ryan being on the GOP ticket in 2012. However, Wisconsin is lily white and blue collar. Look for the cities of Madison and Milwaukee to propel Clinton to victory, but it will be closer than Obama's 7 point win in 2012. Prediction: Clinton by 4

One of the evening's biggest surprises, Wisconsin flipped for the first time since 1984. Depressed inner-city turnout  coupled with Trump winning a number of counties that had long been Democratic for that flipped 10 more electoral votes for the GOP. Actual: Trump by 2

Utah - Another shocker that Utah is even in play. The GOP typically racks up victories of 25% plus. Trump's unpopularity in the Mormon community and conservative local boy turned presidential candidate Evan McMullin are making this closer than expected. Closer between Trump and McMullin that is! Clinton runs a real risk of coming in third place in Utah. Prediction: Trump by 9 over McMullin with McMullin and Clinton in a near tie.

Trump ended up winning easily. Actual: Trump by 19 over Clinton with McMullin losing to Clinton by 9 points to take 3rd place.

Colorado - Closely divided between rural conservative and urban, educated, and liberal regions Colorado seems to be swing state. Yet again, a Hispanic wave is going to sink Trump's chances despite some poll tightening. Prediction: Clinton by 6

A bit closer than I expected... Actual: Clinton by 3

Missouri - This former swing state keeps drifting to the GOP. Clinton will keep it fairly close, but Trump should not have a major problem in holding here. Prediction: Trump by 7

Very easy Trump win here. MO is no longer a swing state. Actual: Trump by 19

The Prof's Home State of Massachusetts

I did well on the calls in Massachusetts at least!

Prediction and Actual

Clinton      65   61
Trump       29   34
Johnston    6    4
Stein           2    1

All incumbent Democratic Congressional Representatives cruise to 20 point plus victories. Only 4 out of the 9 districts have even token opposition.

Massachusetts Ballot Questions

#1 Casino/Racetrack

Yes 33   39
No  67   61

#2 Charter School Expansion

Yes 47   38
No  53   62

#3 Be Nice to Farm Animals

Yes 79   78
No  21   22

#4 Legalization of Marijuana for Recreational Use

Yes 52   54
No  48   46

Other Fun Predictions

Prediction and Actual

Top 5 Clinton States

Massachusetts +36  +27
New York +34  +19
Vermont +34  +29
Maryland +32  +25
California +29  +28

Top 5 Trump States

Oklahoma +26  +37
Alabama +24  +28
Wyoming +21  +48!
Arkansas +21  +27
West Virginia +19  +42

Gary Johnson breaks 10% in his home state on New Mexico and gets nearly 10% in Libertarian-friendly Colorado. Johnson received 9% in NM and 5% in CO.


The House remains Republican albeit with a reduced majority.

The GOP only lost 9 seats, much better than anticipated.

The Senate flips to the Democrats with pickups in IL, WI, PA, NH, and MO. This puts it at 50/50, but the new VP Tim Kaine will be the tiebreaker.

Wrong again - the GOP held the seats in WI, PA, and MO and will have a 52-48 majority. Can we say Unified Republican Government on 1-19-2017...

I will write up a postmortem of the changes that we can expect with a President Trump, a look at the future of both parties, and how to address the faulty public opinion polling.

Onwards to 2018 and 2020!

The Prof

No comments: