06 November 2014

The 2014 mid-terms are now history / a short look at Alaska

I took a long break from this politics blog...

but will be back, using this blog almost daily now for the next years, leading up to the 2016 presidential election.

In about 2-3 months, I will be doing a pretty major analysis of how the pollsters did in 2014. Though most of them picked the correct winner in 8 of the 10 marquee races for the US-Senate, the margins for many, many pollsters were just miserably off, and this time, sharply to the LEFT instead of to the RIGHT.  

Instead of doing finger pointing, which I am sure the pundits are going to do, I will simply wait until the final canvasses come out and then I will do an analysis as I did in February, 2013, for the US-Presidential election (Obama vs. Romney).

But it's not as if I didn't follow the 2014 polls, at least for the epic battle for control of the US-Senate. I did, and essentially blogged the weekly aggregates from RCP here:

2014 battle for control of the US Senate

That study was started on August 11, 2014.

There will be lots of things to look at, but one thing I find especially interesting, in light of electoral history, is what is going on in the great state of Alaska.

Statistically, the state is not as red as it was, for sure.

2000: Bush 43 (R) wins Alaska by +30.95 (+31)
2004: Bush 43 (R) wins Alaska by +25.55 (+25.6) - shift of 5.4 to the Democratic Party
2008: McCain (R) wins Alaska by +21.54 (+21.5) - shift of 4.1 to the Democratic Party
2012: Romney (R) wins Alaska by +13.99 (+14) - shift of 7.5 to the Democratic Party

So, over twelve years, at the national level, the GOP margin in Alaska has shrunk from +31 to +14. The margin, though still a landslide margin similar to Obama's wins in Oregon and Washington State, is less than half of what it once was. Romney's +14 margin in Alaska is the leanest margin for a Republican in a two-way race in this state since 1968, where Humprey really surprised and came dangerously close to Nixon. And those four presidential cycles all overlap the last five Senatorial cycles, including this one, where interesting things have happened since 2002.

A little known piece of history about both Alaska and Hawaii: there were intense fights on the floor of the US Senate about both of these states, because the Democrats at that time were just sure that Hawaii was going to be a "Republican" state and the Republicans of that day were just sure that Alaska was going to be a "Democratic" state, exactly the opposite of how things turned out, according to our current labels for "Democrat" and "Republican". But in the 50's, much of the Democratic Party was far more Conservative than many elements in the Republican party and in the 50's, much of the Republican Party was far more Liberal than many elements in the Democratic Party. This is why BOTH states were admitted to the Union in the same year, with Alaska being admitted on January 3, 1959, the very day that the newly elected 86th congress, with a Democratic majority in the Senate -after Eisenhower's party just lost 13 seats in the 1958 mid-terms, the DEMS also added both Alaska seats and the DEMS moved to a filibuster proof majority of 64-34, which then grew to 65-35 once Hawaii's two Senators were seated - and Hawaii joined the Union on August 21, 1959, therefore giving both states enough time to organize themselves for the next presidential election. In this way, a sort of "balance of power" was kept in the EC, at least in the minds of the politicians of that day, very similar to the admission of Michigan and Arkansas to the Union in 1836-1837.

1990: Ted Stevens (R) wins a seat by +34.04 (+34) - a blowout
1992: Frank Murkowski (R) wins his first Senatorial by +14.64 (+15.6)
1996: Ted Stevens (R) wins re-election by +64.19 (+64.2) - absolute blowout
1998: Frank Murkowski (R) wins a Seat by +54.77 (+55.8) - absolute blowout
2002: Ted Stevens (R) wins re-election by +67.66 (+67.7) - absolute blowout
2004: Lisa Murkowski (R) wins re-election by only +3.03% (+3) against Tony Knowles
2008: Mark Begich (D) unseats Ted Stevens, by only +1.34% (+1.3)
2010: Lisa Murkowsi, as a "write-in", wins re-election, by +4.01 (+4)
2014: Right now,it sure looks as if Sullivan (R) has won and the current statistic is +3.60 (+3.6)

So, after 2002, Alaska went from being an unbreakable +30 points (or way more) solid RED state at the Senatorial level, to a low single-digit win state, now for four cycles in a row. It's also been a minority win state for those four cycles, partly due to Alaska's long tradition of a pretty steep 3rd and 4th party vote.

The Gubernatorial-scene in Alaska has been much more mixed all along:

1998: Tony Knowles (D), wins by 33.01% (+33). Barred from seeking a third term, he ran for Senate against Lisa Murkowski in 2004 and narrowly lost.
2002: Frank Murkowski (who was a sitting Senator) wins by +15.15% (+15.2)
2006: Sarah Palin (R) wins by +7.36% (+7.3)
2010: Sean Parnell, who was Sarah Palin's Lt. Governor and who assumed office after she resigned, wins by +21.39% (+21.4)
2014: Right now, it looks like Independent/Unity Ticket Bill Walker by +1.41 (+1.4)

No matter how you look at it, at all three major levels, two of which are federal, Alaska has gone from being a rock-solid double digit win state pre-Millenium to a single digit-state or sharply reduced double-digit state (presidential cycles) in the new Millenium.

Something is going on with the Alaskan electorate and the historical stats from 12 years back that statement up.

And in one point, I think that former Gov. Sarah Palin was very right: Alaska had indeed by ruled by a few select dynasties for a pretty long time, above all else, the Youngs and the Murkowskis. And the Begiches and the Knowles.

02 October 2014

2016 GE: Hillary Clinton vs. GOP Field, Part V

This is report no. V, a continuation of the Hillary Clinton vs. GOP polling series that I started on March 17, 2013:

Statistikhengst's ELECTORAL POLITICS - 2013 and beyond: Clinton vs. GOP field, 2016 GE, Part I

Here are Part II , Part III and Part IV.

Back in March of 2013, 14 states had been polled. As of August 6th, 2013, it was 21 states. As of now, 28 states have been polled, 27 of which contain presidential match-ups (the lone California poll only has Clinton FAV/UNFAV ratings). Here is how this looks on a map:

As of today, state and national polls together, there have been:

213 polls total, making for 741 presidential matchups.

Hillary Clinton has won 585 of those 741 matchups (78.95%)
The GOP field of various candidates have won 141 of those matchups (19.03%) and there have been 15 mathematical ties (2.02%)

So, in spite of the fact that the number of polls has more than doubled in the last ten months and the number of matchups has more than tripled since then, the actual statistic of wins for Hillary Clinton has remained very stable.  I would like to show this with three charts, all self-explanatory:

Report date no. State polls no. state matchups Clinton GOP ties margin
17 Mar 2013 15 34 82.35% 14.71% 2.94% +67.64
06 Aug 2013 40 102 73.53% 21.57% 4.90% +51.96
16 Nov 2013 58 158 75.95% 21.51% 2.54% +54.44
11 May 2014 115 386 75.65% 21.51% 2.85% +54.14
02 Oct 2014 166 602 74.75% 22.76% 2.49% +51.99

Report date no. Nat'l polls no. Nat'l matchups Clinton GOP ties margin
17 Mar 2013 5 11 100.00% 0.00% 0.00% +100
06 Aug 2013 21 38 100.00% 0.00% 0.00% +100
16 Nov 2013 25 51 98.04% 1.96% 0.00% +96.08
11 May 2014 41 122 96.72% 3.28% 0.00% +93.44
02 Oct 2014 47 139 97.12% 2.88% 0.00% +94.24

Report date no. all polls no. all matchups Clinton GOP ties margin
17 Mar 2013 20 45 86.67% 15.55% 0.22% +71.12
06 Aug 2013 61 140 80.71% 15.72% 3.57% +64.99
16 Nov 2013 83 209 81.34% 16.75% 1.91% +64.59
11 May 2014 156 508 80.71% 17.13% 2.17% +63.58
02 Oct 2014 213 741 78.95% 19.03% 2.02% +59.92

The column on the left of each table is the report date. One table is only for the state polls/matchups, the second table is for the national polls/matchups and the third table is for all polls combined. Look at how very stabile those numbers have remained over more than 1.5 years.

Here is an exact table, by state, with the number of polls, matchups, and who won:

State No. of polls No. of Matchups Clinton wins GOP wins Ties
National 47 139 135 4 0
AK 6 30 6 24 0
AR 4 15 5 10 0
AZ 1 5 4 1
CA 1 0 0 0 0
CO 10 36 21 13 2
FL 15 51 51 0 0
GA 2 9 7 1 1
IA 14 59 52 4 3
KS 4 16 1 15 0
KY 5 14 6 7 1
LA 7 27 10 14 3
ME 1 4 4 0 0
MI 9 30 30 0 0
MN 2 9 9 0 0
MS 2 9 0 8 1
MT 3 8 0 7 1
NH 9 26 25 1 0
NJ 9 18 17 1 0
NM 1 6 6 0 0
NY 7 11 11 0 0
NC 12 47 41 6 0
OH 6 33 32 0 1
OR 1 5 5 0 0
PA 8 34 33 1 0
TX 4 18 5 13 0
VA 15 45 43 1 1
WV 1 5 0 5 0
WI 5 24 23 0 1
WY 1 5 0 5 0
Latino 1 3 3 0 0
TOTAL state polls 166 602 450 137 15
TOTAL state and national 213 741 585 141 15
% state polls only

74,75% 22,76% 2,49%
% state and nat'l together

78,95% 19,03% 2,02%

All of the polling values are here in this EXCEL DOCUMENT. I encourage all readers to go to this link and use the tabs at the bottom to see their state first and peruse as much as they wish.

Here is a screenshot of an example of  the layout of this table, using the state of Florida as an example:

You will notice that the polls are numbered in reverse chronological order, with the latest poll at the top. All polls are linked, so you can click on the link for every single poll and see the results for yourself. The release date of the poll (European dating system) is then following by the survey size and MoE (Margin of Error). Then, if there are any FAV/UNFAV numbers or DEM nomination figures, they come next. I am not averaging polls yet.

The Presidential matchups follow in this order:

Clinton vs. Christie
Clinton vs. Paul
(at the point the screenshot shows no more, you would have to scoll out at the excel table to see the rest)
Clinton vs. Ryan
Clinton vs. Cruz
Clinton vs. Bush, J.
Clinton vs. Rubio
Clinton vs. Huckabee

And then, any other GOP candidates that have been polled in that particular state.

The numbers are COLOR CODED (blue = DEM / red = GOP) and bolded = the higher value. Wherever you see values in italic, that means a mathematical tie.

The following 31 pollsters have polled Hillary vs. GOP matchups thus far:

NBC (Princeton)
Marist / McClatchy
The Field Poll
Gravis (R)
Harper (R)
Purple Strategies
Marquette University Poll
Latino Decisions
The Arkansas Poll
The Polling Company
Conservative Intel
Loras College
Civitas (R)

Here is a map of those states that have been polled, colored by the winner of the majority of the match-ups. In the case of Colorado and Kentucky it's really quite close, so I am leaving both states green for now.

This map is NOT a prediction map. It only shows who has won the majority of matchups. And interestingly enough, it has not changed even one bit since the last report - from May, 2014.

What to make of all of this?

Well, it's still early, but the trend we have seen all through 2013 has continued through 2014, namely, that Hillary Clinton is demonstrably ahead in the battleground states that have decided the last 6 elections cycles. Not only is she ahead in those states, she is decisively ahead.  And in states where the GOP is winning, the aggregate margins are reduced.

1.) The Quntifecta: Florida (29 EV), Pennsylvania (20 EV) Ohio (18 EV), North Carolina (15 EV) and Virginia (13 EV). Total EV: 95

In every one of those 5 states, Hillary is considerably ahead of her potential GOP opponents.

In Virginia, one of the most polled states thus far, there have been 15 polls with 45 matchups. Hillary has won 43 of them, there was one tie and Christie (R) won one matchup in a poll from 2013 that I missed and just included in the statistic (Washington Beacon Poll). All said and told, Hillary is at between +4 and +14 against GOP candidates, all margins larger than Obama's 2012 win. 

In Florida, the third most polled state thus far, there have been 15 polls with 51 matchups and Hillary has won every single matchup, with the majority of the margins in double-digits. No Democratic candidate that I know of has scored these type of margins in Florida in modern polling history. Not even Bill Clinton scored these kinds of numbers.

The kind of polling we are seeing out of Florida and Virginia should be the no. 1 warning sign for the GOP.

In North Carolina,  there have been 12 polls and 47 matchups, of which Hillary Clinton has won 41. A strike against this data is that all of the polls except one have been from PPP (D) and I would much prefer to see a broad base of pollsters, as is the case with Virginia and Florida. The margins are also smaller, but consistent. PPP (D), which is based in North Carolina, nailed the polling in 2008, but called a tie in NC in 2012, where Romney won by +2.04%, so PPP (D) technically missed the NC call in 2012. Alone the fact that this state is still unbelievably competitive is a bad sign for the GOP.

In Ohio, somewhat less polled than the others, there have been 6 polls and 33 matchups, of which Hillary has won 32. And in 14 of those 32 wins, Hillary is winning with double digit margins over her opponents. The last time a Democrat won Ohio with more than a single-digit margin: LBJ, 1964. Before that? FDR, 1936. Both of those elections were massive blowout elections for the Democratic Party. Bill Clinton barely won Ohio in 1992, but he won it by +6 in 1996. Pres. Obama won Ohio by +4.6 in 2008 and by +3 in 2012. There has not been a poll of Ohio since 1988 to show a candidate of any party in double digits. This is an extremely important data point to remember. Again, it's not just one poll or one matchup where she is winning with double digits: it's in 14 matchups spread out over 4 polls.

In Pennsylvania, also one of the less polled states, there have been 8 polls and 34 matchups, Hillary has won 33 of them. And, similar to Ohio, 23 of those 33 wins are double-digits wins. Here, Christie still does the best of the GOP field. Pennsylvania is technically no longer a swing state, having not swung for Pres. George W. Bush (43) in either 2000 or 2004. But it has been a battleground in the last 4 elections, to be sure.

So, of the Quintifecta, we have four of five states (VA, OH, FL, PA) where, if the double-digit margins hold like we are seeing right now, those states would not even be true battlegrounds on election day 2016.

2.) Clinton is showing considerable strength in some of the so-called "Clinton 6 states". 

The "Clinton 6", as I call them, are the six southern -or- Mason-Dixon states that Bill Clinton won in both 1992 and 1996, states that Pres. Obama never won. They are: WV, KY, MO, AR, LA and TN.  I did a write up over this phenomenon in November 2012.

Four of those states have been polled to-date: WV, KY, LA and AR.

There is only one poll out of West Virginia, but it confirms that the massive Romney landslide victory of 2012 is likely to hold for any Republican in 2016. 

However, in Kentucky, it really could be a horserace (Bill Clinton won the Bluegrass state twice), and in Arkansas (the former home-state of the Clintons), we now have a fourth poll showing a tight single-digit race, except against fellow home-stater Mike Huckabee. 

To note, Mitt Romney won West Virginia by +27 and Arkansas by +24 and also made little-known electoral history in 2012 as being the first Republican ever for whom both of these erstwhile Democratic bastions were called for a Republican immediately at poll closing time on election night, but the Clinton polling here is divergent: she is losing in WV but winning somewhat in AR. However, she is under 50 in AR, there are lots of undecideds and things could change again. I have personally been thinking that Hillary would have better chances in Missouri, a state where Obama and McCain practically tied in 2008, but Romney won handily in 2012, than in Arkansas. Wait and see.

One of the advantages of such an interesting 2014 mid-term election season is that some states that are rarely polled for presidential elections are getting attention right now, including the great state of Louisiana, which has now been polled 7 times and with 27 matchups, of which Hillary was won 10 matchups and the GOP candidates have won 17. More importantly, in most all matchups, it is a close single-digit race. In the last PPP poll of Louisiana, the margins went from Hillary +4 over Christie to Huckabee +7 over Hillary.

3.) Individual states that had been battlegrounds in 2000, 2004 and to some extent, in 2012, look very solid for Clinton: 

Wisconsin, Minnesota, Michigan, New Mexico and New Hampshire.  


in the West, also part of a very consistent pattern, Clinton is struggling mightily in Colorado, she has the entire time. If there is a state that the GOP has the best chances of regaining from 2008-2012, it is probably the Rocky Mountain State. Wait and see. There have now been 10 polls of Colorado and 36 matchups, of which Hillary has won 21, the GOP has won 13 and there were two mathematical ties. Then again, polling in Colorado was way off to the Right in 2008, 2010 and 2012. Why? The Latino vote was grossly underestimated and then also grossly miscalculated. Wait and see what happens in 2016 in the Rocky Mountain State.

 It also appears to be close in Iowa, but Clinton is winning.

In the state of New York, an expected blue state, Clinton has margins upwards of +40. The last (and only) time a Democrat won NY with circa +40? LBJ, 1964. 

Likewise, the GOP is easily winning Wyoming, an expected red state, by an average of about +27. George W. Bush (43) won Wyoming with over +40 both times.

Now, you might say: "Why quote NY and WY? NY is going to go blue and WY is going to go red!" And I will say: "a rising tide lifts all boats".

Expanded margins for Clinton in expected blue states and suppressed losing margins for her in expected RED states would point to a national win for her.

Speaking of red states, here is a surprise red state for you: KANSAS

In the great state of Kansas, there have now been 4 polls and 16 matchups, of which the GOP has won 15 and Hillary has won 1, but the margins are interesting:

Nr. KANSAS Date Sample MoE Clinton Christie Und/Oth Mar.
Clinton Paul Und/Oth Mar.
Clinton Ryan Und/Oth Mar.
Clinton Cruz Und/Oth Mar.
Clinton Bush, Jeb Und/Oth Mar.
Clinton Rubio Und/Oth Mar.
Clinton Huckabee Und/Oth Mar.

AVERAGE (one week): N/A N/A N/A #WERT! #WERT! #WERT! #WERT!


4 PPP (D) 16.09.14 1328 LV +/-2.7 40 42 18 2
41 43 16 2

100 0
44 41 15 -3
40 46 14 6

100 0
42 45 13 3
3 PPP (D) 22.08.14 903 LV +/-3.3 38 42 20 4
41 45 14 4

100 0
42 43 15 1
39 45 16 6

100 0
41 46 13 5
2 PPP (D) 24.02.14 693 RV +/-3.7 39 46 15 7
41 48 11 7

100 0

100 0
41 50 9 9

100 0
42 49 9 7
1 PPP (D) 27.02.13 1229 RV +/-2.8

100 0

100 0
43 50 7 7

100 0

100 0
42 47 11 5

100 0

Indeed, in Kansas, it is a narrow single-digit race right now. Is it likely to stay that way? Probably not. But Kansas is supposed to be a +20 to +30 RED state. Right now, were these numbers to hold, Kansas would be a battleground state.

4.) Speaking of national win: national polling is overwhelmingly pointing to a large Clinton win:

There have been 47 national polls, with 139 matchups. Hillary Clinton has won 135 (97.12%) of those matchups.  Go click on the national polling tab and see for yourself how many of those margins are double digit margins.

Now, varying from pollster to pollster and month to month, one GOPer may come closer to Clinton than the others, but she beats them all, consistently. And this data is coming from many various and independent-from-each-other pollsters. Rasmussen, a Right-Leaning outfit, recently put out a poll showing Hillary Clinton with a +13 margin over Jeb Bush nationally.

5.) My wish-list: 

there are some states for which we have no data right now and it would be very enlightening to get some data from such states as, for instance, Missouri, which used to be quite a bellwether state in our Union (until 2008), and Nevada, which was considered a battleground in 2008 and in 2012. Also, from the 9.25 Obama pick-ups from 2008, it would be good to see data from Indiana, which went for Obama by +1.03% and then swung back to the Republicans with Romney and was won by +10.20 in 2012. However, Indiana has been traditionally a solid +20 RED state, so indeed it would be interesting to see how Hillary does in rural areas in such a state. My wish would also to see more polling outside of PPP (D) and Quinnipiac. Also, Obama won Nebraska-02 in 2008, it would be interesting to see how Hillary is doing in the congressional district this time around.


Facit: it's Hillary's election to have, if she wants it. If she declares (and I am more than reasonably sure she will declare), then she is the prohibitive front runner both for her party's nomination and also against any and all comers from the GOP.

Factors working against the GOP are the fact that Obama literally cemented Virginia into the Democratic column by winning the state not just once, but twice, against all conventional wisdom. Another factor against the GOP was the wild swing of Cuban-American voters in Florida from the GOP to Obama in 2012. The assumption is that this was due to the ongoing immigration debate. Another factor is the name "Clinton", where both Hillary and former President Bill Clinton are extremely well-known to the American public and in spite of the Lewinski scandal, Bill Clinton is still quite respected among the American public.

If Hillary only wins the states that were common DEM states between Bill Clinton and Barack Obama, then she is already at 290 EV (shown in the Clinton 6 link above, here is the link again). I see a real possibility for Hillary Clinton to go slightly over 400 EV in 2016.

Those are the current polling statistics. Update in January/February 2015. Starting with the next analysis, I will be publishing electoral maps for Hillary against potential GOP opponents by name.

Full disclosure: I am a Clinton supporter, but were the numbers to be showing the GOP running away with this race, I would report it with exactly the same veracity.