Saturday, October 24, 2020

I predict the Democrats will take the Senate, holding 51 Senate seats when the new Congress convenes on 1/3/2021.

I predict the Democrats will lose a seat in Alabama and win GOP held seats in Arizona, Colorado, Maine, North Carolina and Iowa.  I predict Lindsey Graham will be reelected in South Carolina for the GOP by a tiny margin, perhaps in a recount.  I predict the GOP will eventually hold both Senate seats up for election in Georgia (though the Democrats COULD win 1 or even both on a BIG Biden blowout win).  There are other Democrat opportunities for pickups, notably Montana, but absent a Biden win by more than 10 points nationally I don't expect any of these to come to fruition.  Finally, the GOP has 1 other longshot pickup opportunity in Michigan.  I expect the Dem to win by about 5 points, not a sterling performance, but enough to keep his seat.

The Democrats currently have 47 seats (counting the two "independents" that caucus with the democrats, Bernie Sanders and Angus King of Maine).  Assuming Biden wins the election (as I strongly predicted in my previous post), the democrats need 50 Senators to control the Senate (with newly elected Vice President Kamala Harris breaking the tie).  Doug Jones of Alabama will lose.  Yes the polls are close.  Yes he will do much better than Biden in Alabama.  But Alabama is too red to elect a democrat unless, say, a child molester running against him.  So the Democrats will "start" with 46 seats and need 4 to take control (again assuming Biden wins). 


Prediction: Democrat Mark Kelly by 8.

Fundamentally, Arizona should be a Biden state (but not by much at all!), and the democrat should run well ahead of Biden.  Easy democratic pickup. Predict the democrat wins by 8.

In Arizona, Nate Silver's model, at  predicts the Arizona Senate vote will be 52.8-47.2, or the democrat, Mark Kelly, by 5.6 points.  I would definitely take the over on that.  Biden will likely (but not definitely) win Arizona, the state hasn't warmed to incumbent Republican Martha McSally, and Kelly is a GREAT candidate (former astronaut and husband of former Congresswoman Gabrielle Giffords, who was shot in the head at point blank range some years ago and survived.


Prediction: Democrat John Hickenlooper by 9

Colorado will be an easy pickup for the Democrats.  CO is a blue state at this point; I have Biden winning it by 12.5 (Nate Silver has CO going for Biden by an expected 11.5 points).  Voters rarely split tickets these days.  The Colorado Senator is linked to Trump.  The democrats are running a mildly popular former governor.  John Hickenlooper is not a natural politician, but he's acceptable.  Nate Silver's model has Hickenlooper predicted to win by 7.6 points.  I would again take the over, but not by much.  I used to think Hickenlooper would outrun Biden; I'm not so sure of that now.  But it won't matter.


Prediction: Democrat Sara Gideon by 6

Susan Collins, aka Concerned Susan for Democrats that follow this stuff closely, will lose.  She used to have a LOT of crossover appeal, with Democrats voting for her in significant numbers in past elections.  Not this time.  Nate's model has her losing by 2.6 points to Sara Gideon, the state House Speaker. I would yet again take the over, as I expect Gideon to win by about 6.

If I only expect a 6 point win how can I be so very confident?  A very fair question.  Especially since Maine is an elastic state (meaning it has relatively more voters than average that will change their minds).  It would take a while to explain, but the short version is that Gideon is up in virtually every poll and more importantly Biden will win Maine big.  Hillary C only won Maine by 3, whereas I have Biden winning it by 12!  Collins will outperform Trump for sure in this light blue state, but not by enough to make up a 12 point stagger.  

North Carolina:

Prediction: Democrat Cal Cunningham by 3

If the 3 above races go as I predict, North Carolina would be the 4th seat that Democrats would need to take the Senate.  I predict that the Democrats will win this seat, although we reach the first seat where I'm not super sure.  I have Biden winning North Carolina by 1.  While I'd way rather be Biden than Trump in NC, I'm obviously not sure about Biden winning it.  Democrat Cunningham should outrun Biden.  This was a near certainty until Cunningham became embroiled in a sexting scandal.  Most important Senate race for determining Senate control (the most likely Senate tipping point if you will) and the GOP candidate gets COVID and the Democrat has a sexting scandal.  So 2020.

Anyway, we should know the winner of NC, and this seat, on election night, unless one or both are VERY close, as NC will have its ballots counted very quickly (as will Florida).


Prediction: Democrat Theresa Greenfield by 1

I have Trump winning Iowa by about 1.  I do expect Greenfield to outrun Biden, but NOT by much.  So while I am predicting Greenfield to win, the real answer is: (1) I don't know who is going to win this race; and (2) you don't know either.  How Biden does nationally will likely decide this race.  If Biden wins in a blowout, by say 10 points nationally, the Republican, Joni Earnst, is toast.  If Biden wins by only 6, I would predict that Ernst survives and holds the seat for the GOP.  At my 8 point prediction this race is close but SLIGHTLY favors the democrat.  I note that Nate Silver's model forecasts the election as a 1 point win for the Democrats.  So I'm basically cheating.  Don't tell anyone please.


Prediction: GOP Senator Perdue by about 1.5 points.  

Prediction 2: The other seat is CERTAIN to go to a runoff election IN JANUARY, likely with 1 Democrat and 1 Republican.  I think the Democrat will be a modest underdog, though I am very uncertain about that.  Usually Democrats lose runoffs, but if the Biden rout that I predict does come to pass, there is a real chance of a dispirited GOP base in Georgia simply not bothering to show up for a runoff while a still enthused Democrat base does.

I have Biden winning Georgia by 0.8 points.  Wouldn't be surprised by 3 points either way on that prediction.  Crucially, Perdue is VERY likely to outrun Trump. Hence my prediction that he hangs on.  But on a big Biden night, Perdue could well lose.  The problem Democrats have here is that it is Trump that is causing Georgia to be so close.  There is no objection to Perdue that I am aware of among the state GOP and right leaning independents.  Hence he outruns Trump.  Hence he probably holds on.

Wednesday, October 07, 2020

 I predict Biden will beat Trump by 8 points, and win the electoral college in a landslide.

I predict Biden beats Trump by 8 points in the popular vote, 53.2-45.2. Obviously I can’t predict with precision, but that is my working hypothesis.  I predict Biden gets 369 electoral votes, 4 more than Obama did in his 2008 landslide. 

Of the swing states, I have Biden winning practically all of them: Michigan, Wisconsin, Pennsylvania, Florida, Ohio (probably), Georgia, North Carolina, and Arizona.  Note that assuming that Biden wins every state Hillary Clinton won, which is very likely), the 3 Midwestern states alone (Pennsylvania, Wisconsin, and Michigan) are enough to win Biden the electoral college and thus the presidency.

The map I am predicting is just below.  

I expect Biden to do VERY well in the Midwest (Pennsylvania, Wisconsin and Michigan) and give him slightly better than 50-50 to win Ohio (which Trump won by 8 in 2016!)  I expect Biden to do only ok in North Carolina (15 electoral votes) and Florida (a whopping 29) but because he’s winning by 8 nationally, I expect him to carry both states, and Florida by 2 points (practically a landslide for Florida!) 

Clinton beat Trump by 2.1 points (note: I mean Hillary Clinton throughout; I do not discuss Bill at all).  Given that I expect Biden to win by 8 points, that means I expect the nation to be 6 points bluer than in 2016.  Some states will move more than 6 points towards Biden (the Midwest in particular!) and some states will move less (Florida, North Carolina, Georgia (I think) Arizona (I think).  However, Trump won each of the 3 midwestern states, Florida, North Carolina and Arizona by less than 4 points (significantly less in the case of Florida and the 3 Midwestern states).  Thus, if my popular vote prediction is correct, he should win every swing state I named except perhaps Ohio which should be very close.

Detailed discussion of the swing states (and a few non swing states below).

Why should anyone read my election predictions?  Well, in June 2008, I put up a post here,

in which I predicted an Obama landslide 8-point win and made predictions for the all the swing states and some others.  My predictions were spectacular!!!  See the summary of the results of my predictions here.

My 2012 predictions were even better, however 2012 was a MUCH easier election to predict.  I missed only one state, Florida, which Obama only won by 0.9 points.  The post I made, was the night before the election.  You'll have to trust me that I was predicting the exact same thing for months.

My 2016 predictions were TERRIBLE.  I never put up a blog post (fortunately?), but I had Clinton winning a mess of states that Trump won.  Just a horror show.  Hey, 2/3 aint bad. 

With respect to 2020, I had been expecting a closer popular vote until just several days ago.  Had I written this a month ago, I would have had Biden winning by about 6.  Trump's bullying debate performance, and his contracting Covid (as well as so many others around him), after all of his denial and misinformation about it, seems to have hurt him significantly at least in the short run.  I expect that to fade somewhat, but not entirely.

Note that I am not factoring in any shenanigans/plain old theft on the part of the GOP, nor the certainty of a lot of mail in ballots being rejected.  While a real risk (and in the case of mail ballots being rejected a certainty): (1) I’m not that worried about GOP shenanigans; (2) I have no edge in understanding the risk; and (3) while mail in ballots will be rejected much more often than in person votes, they also make it easier to vote while increasing turnout. 

Onto the meat of my predictions!


2008: Obama + 7.2

2012: Obama + 3.9

2016: Clinton + 2.1

2018 House popular vote (the aggregate of all votes nationwide in the 435 US House elections): Dems + 8 (close enough)

2020 Real Clear Politics ("RCP") AVG as of 10/7/2020: Biden v. Trump: Biden + 9.1

Predicted 2020 popular vote Biden + 8

Arizona (11 electoral votes): 

Predicted 2020 lean: Red + 5 (moving .6 points towards blue relative to the nation)

Predicted 2020 Winner: Biden by 3

By predicted lean, I mean where I expect the state to be relative to the 8 point predicted popular vote win for Biden.  I predict Arizona to be 5 points more Republican than the nation, hence my predicted 3-point win for Biden (8 point popular vote win -5 point Republican lean = 3 point win).

2008: McCain + 8.5 (red + 15.7) (3rd party 1.6%)   (2.32 m votes cast) 

2012: Romney + 9 (Romney a great fit) (red + 12.9) (3rd party 2%) (2.307 m votes cast)

2016: Trump + 3.5 (Trump a bad fit) (red + 5.6) (3rd party 7.5%) (2.605 m votes cast) 

Final 2016 RCP average GOP + 4.  Deviation Dem + .5

2018 Senate: Dem + 2.35 (red + 5.65) (3rd party 2.4%) (2.384 m votes cast) 

RCP average on 10/7/2020: Biden + 3.4 (Red + 5.7)

Comments on Arizona: Arizona has a lot of seniors, a group which Biden is greatly outperforming Clinton.  While Biden is underperforming Clinton with Hispanics nationally, he does not appear to be doing so in Arizona according to the polls.  Arizona also has a fair number of suburbs, where the GOP has badly underperformed 2016, both in actual elections and in polls.

The 2018 Arizona Senate race, which Sinema, the democrat, won by 2.35 points is instructive.

Arizona Senate race: GOP Senator Martha McSally is toast.  I predict she loses by at least 8.  This would give democrat Mark Kelly the seat for only 2 years, as this is a special election

Colorado (9 electoral votes: 

Predicted 2020 lean: Blue + 4.5 (moving 1.8 points towards blue relative to the nation)

Predicted 2020 Winner: Biden by 12.5 (!!!)

2008: Obama + 9 (Blue + 1.8) 3rd party 1.7%  (2.4 m votes cast) 

2012: Obama + 5.4 (Blue + 1.5) (3rd party 2.4%) (2.57 m votes cast)

2016: Clinton + 4.8 (Blue + 2.7) (3rd party 8.6%) (2.78  m votes cast) 

Final 2016 RCP average Dem + 3   Deviation Dem + .1.9

RCP average on 10/7/2020:  Biden + 10 (blue + 2.5) (ONE POLL!)

Comments on Colorado:

Colorado has a ton of people with college degrees (which favors Biden) and is quite suburban (which favors Biden, even as compared with 2016).  Although I see a Biden blowout in Colorado in 2020, I could easily see it being competitive in 2024 depending on the national environment, and the type of candidate and campaign run by the GOP.  Colorado is a blue state now, but its not so blue that it is out of reach for the GOP.  It is about as blue as Florida is red, for reference, and a Democrat can of course win Florida in a good year (Obama won it twice).  If I were a GOP primary voter in 2024 concerned about winning above all else, Colorado would one of the top two or three states I would have in mind (Florida and Pennsylvania round out the list).

Colorado Senate race: Incumbent GOP Senator Cory Gardner is toast.  He should lose by low double digits to former Colorado governor John Hickenlooper.  Gardner isn’t especially noxious or unpopular, but he’s tied to Trump, and that is enough for him to lose in this light blue state with a very unpopular (in Colorado) Trump on the ballot.  Bye bye Cory!

Florida (29 electoral votes !!)

Predicted 2020 lean: Red + 6 (moving red by 2.7 points relative to the nation) 

Predicted 2020 Winner: Biden by 2 (!)

2008: Obama + 2.7 (red + 4.5) (3rd party 0.9%) (8.411 m votes cast)  

2012: Obama + 0.9% (red + 3) (3rd party .87%) (8.474 m votes cast) 

2016: Trump + 1.2 (red + 3.3) (3rd party 3.16%) 9.40 m votes cast) 

Final 2016 RCP average GOP + 0.2 -- Deviation GOP + 1

2018 Senate: GOP + 0.13 (red + 7.87) (3rd party almost zero) (8.189 m votes cast)

RCP average on 10/7/2020: Biden + 3.5 (red + 5.6)

Comments on Florida:

Florida is a rare swing state where Biden hasn’t polled all that well.  He’s up 3.5 points in the average as of now, but Biden has not pulled away at all in Florida. Democrats typically underperform their polls in Florida and did so in 2018.  That is a real concern with respect to who wins Florida.  Biden not only could underperform his polls, but SHOULD, and I predict will.  As of now, he has a small margin of error.  That could change.

There are a ton of southern whites, who if anything are more Republican than in 2016.  The large Cuban community still favors the GOP (but not as much as did 20 years ago).  These voters have not moved for Biden in the polls as of yet, and likely will not.  There are a LOT of seniors, and while Trump hasn’t lost that much ground among seniors in Florida compared to 2016, he has lost some ground, and that should tip the balance.  I am not confident Biden will win by as much as 2, it could easily be less.  But it should be more than the 0.5% margin for which Florida law mandates a recount.  Obviously, Trump can sue to try and invalidate the massive numbers of mail in ballots.  If such a suit were even partially successful, Trump would VERY likely win Florida, and win the election.  This is most unlikely, in my opinion.

For those understandably worried/focused on possible GOP shenanigans or lawsuits causing mail in ballots not to count, watch Florida very closely on election night.  Florida is used to huge numbers of mail in votes, and begins to count them a few weeks BEFORE election day (unlike Pennsylvania and some other states which are, amazingly, barred by law from counting mail in votes before election day regardless of when they arrive).  The results in Florida on election night should be quite close to the final results.  If Florida is called for Biden, its hail to the Chief (for Biden!)  If Florida is too close to call, Biden can and should still win if all the mail in votes in the Midwest are counted.  If Florida is called for TRUMP (would surprise me), Biden won’t win the popular vote by 8, and there is at least some suspense (perhaps a lot) over who will win the electoral college.

Georgia 16 electoral votes

Predicted 2020 lean: Red + 7.2. (not moving relative to the nation)

Predicted 2020 Winner: Biden by 0.8

2008: GOP + 5.2 (red + 12.4) (3rd party 1%) (3.932 m votes cast) 

2012: GOP + 7.8 (red + 11.7) (3rd party 1.2%) (3.900 m votes cast) 

2016: Trump + 5.1 (red + 7.2) (3rd party 4.2%) (4.141 m votes cast) 

Final 2016 RCP average GOP + 4.8.  Deviation GOP + 0.3

RCP average on 10/7/2020: Biden by 0.3 (red + 8.8)

Comments on Georgia:

Georgia has lots of white southerners (which favors Trump heavily, as these voters have not changed their preferences much if at all compared to 2016).  However, there are a LOT of black voters, and there is real room for increased turnout as compared to 2016, which would be a huge deal.  I note that one of the two Senate races features a black candidate, which could marginally drive up turnout among black voters.

Finally, Georgia has a lot of affluent white suburban voters, which have turned against the GOP since 2016.  That said, many of these voters already did not vote Trump in 2016.  Adding all this up, I don’t expect much movement relative to the nation.  However, Biden winning the popular vote by 8 points would be too much to overcome.

Georgia Senate races:

Georgia has 2 Senate races, one of them a “jungle primary” election which culminates in a January runoff.  This is a special election which gives the winner only 2 more years.  The other Senate race is a normal Senate race.  If Biden wins by 8, both are too close to call.  FORCED to predict, I’d predict incumbent GOP Senator Perdue hangs on by his fingernails while a democrat makes the runoff, where he’d be an underdog.  So I tentatively predict the GOP hangs onto both of these Senate seats.

Iowa (6 electoral votes)

Predicted 2020 lean: Red + 9 (moving blue by 2.5 relative to the nation)

Predicted 2020 Winner: Trump by 1

2008: Obama + 9.54 (blue + 2.44) (3rd party 1.62%) (1.544 m votes cast) 

2012: Obama + 5.81 (blue + 1.9) (3rd party 1.83%) (1.582 m votes cast) 

2016: Trump + 9.41 (red + 11.51) (3rd party 7.11%!!!!!!) (1.566 m votes cast)

Final 2016 RCP average GOP + 3.  Deviation GOP + 6.5

RCP average on 10/7/2020: Biden + 0.5 (red + 8.6)

Comments on Iowa:

I thought I would end up predicting that Biden wins Iowa.  But when I ran the numbers, I ended up with a prediction of Trump, and I’m not going to let my (flawed) pundit ability overturn the results that my best guess at the numbers.

Iowa is lily white, and has a ton of white non college graduates, which were Trump’s most important strong cohort by far.  Biden is greatly outperforming Clinton among this group, which is why the Midwest looks so terrible for Trump right now.  Iowa is, I think, red enough with Trump on the ballot for him to hang on even while losing nationally by 8 points.  I am not confident in this prediction, however. Biden could win a squeaker here.

Iowa Senate race: Incumbent Republican Joni Ernst is in a tight race.  The RCP average right now is an astounding Greenfield (democratic challenger) by 5!  I don’t believe that for a second.  I tentatively predict that the democratic challenger wins, but I have little confidence in this prediction.  A titanic amount of money has been spent on this race by Iowa standards.

Maine (2 electoral votes statewide, 1 each for the 2 congressional districts)

Predicted 2020 lean: blue + 4 (moving blue by 3.1 relative to the nation) 

Predicted 2020 Winner: Biden by 12 (real downside risk to this forecast, but Biden is safe to win Maine comfortably) He will win at least 3 of the 4 electoral votes, and is a modest favorite to win the 4th

2008: Obama + 17.3 (blue + 10.1) (3rd party 1.9%) (731,163 votes cast) 

2012: Obama + 15.29% (blue+ 11.5) (3rd party 2.7%) (724,758 votes cast) 

2016: Clinton + 3 (blue+ 0.9) (3rd party 7.4%!!!) (771,892 votes cast) 

Final 2016 RCP average DEM +5 4.  Deviation GOP + 1.6  

RCP average on 10/7/2020: Biden + 12.8 (Blue + 3.7)

Comments on Maine:

Maine has a lot of white non college degree voters, the group that swung so hard for Trump in 2016.  A few months ago, I expected Maine to be fairly close this time, but the recent polls predicting a Biden blowout have convinced me otherwise. 

Maine Senate race: Incumbent GOP Senator Susan Collins, aka Concerned Susan, is very likely to lose if Biden carries Maine by double digits.  If Trump keeps Maine a lot closer than I think, the Senate race could be a photo finish.  I absolutely expect Collins to lose and would be very surprised if she pulled the race out.

Michigan (16 electoral votes)

Predicted 2020 lean: Exactly purple-- even with popular vote (moving blue by 2.3 relative to the nation)

Predicted 2020 Winner: Biden by 8

2008: Obama + 16.4 (blue + 9.2) (3rd party 1.8%) (5.010 m votes cast) 

2012: Obama + 9.5 (blue+ 5.6) (3rd party 1.08%) (4.731 m votes cast) 

2016: Trump +.023 (red + 2.3) (3rd party 5.3!!!) (4.799 m votes cast) 

Final 2016 RCP average Dem + 3.4 Deviation GOP + 3.7  

2018 Senate: Dem + 6.5 (red + 1.5) (3rd party .59%) 4.237 m votes cast) 

RCP average on 10/7/2020: Biden +6.2 (red +2.9)  

Comments on Michigan:

Michigan was Trump’s biggest surprise in 2016.  There are a lot of white non college voters, who strongly favored Trump in 2016.  However, these voters shifted hard to the democrats in the Midwest in 2018 (though the GOP still won them, just by much less than in 2016).  The polls say this election will be a LOT more like 2018 than 2016.  Hence Biden’s strong performance in the polls in the Midwest.  I’d be ASTOUNDED if Trump pulled off the upset here.  If Trump does somehow, some way, win Michigan, he’ll very likely be reelected.

Michigan Senate race: I’d be shocked if incumbent Democrat Peters didn’t win.  That said, the GOP candidate, John James, is a strong candidate.  If the popular vote is VERY close (wildly unlikely imho), James could pull off the upset.

Minnesota (10 electoral votes)

Predicted 2020 lean: blue + 2 (moving blue by 2.6 relative to the nation) 

Predicted 2020 Winner: Biden by 10

2008: Obama + 10.4 (blue + 2.45) (3rd party 2.12%) (2.91 m votes cast)

2012: Obama + 7.7 (blue+ 3.8) (3rd party 2.4%) (2.937 m votes cast) 

2016: Clinton + 1.52 (red + 0.62) (3rd party 8.5%!!!) (2.945 m votes cast) 

Final 2016 RCP Dem + 6 Deviation GOP + 3.5

2018 Senate: Dem + 24.1 (blue + 16.1) (3rd party 3.45%) 2.597 m votes cast) (Amy Klobuchar is a VERY strong candidate—Minnesota is NOT that blue anymore)

RCP average on 10/7/2020: Biden + 9.4 (blue + 0.3) 

Comments on Minnesota:

Earlier in this cycle I and others though Minnesota would be at least vaguely competitive.  It won’t.

New Hampshire (4 electoral votes)

Predicted 2020 lean: Exactly purple, even with the nation (moving blue by 1.73 points)  

Predicted 2020 Winner: Biden by 8

2008: Obama + 9.65 (blue + 3.4) (3rd party .89%) (707,611 votes cast) 

2012: Obama + 5.6 (blue+ 1.7) (3rd party 1.6%) (710,972 votes cast) 

2016: Clinton + 0.37 (red + 1.73) (3rd party 5.13%!!!) (743,117 votes cast)   

Final 2016 RCP average Dem + 0.6 Deviation GOP + 0.3    

RCP average on 10/7/2020: Biden + 8.4 (red + 0.7)     

Comments on New Hampshire: New Hampshire should absolutely be competitive in 2024 and beyond.

Nevada (6 electoral votes)

Predicted 2020 lean: Blue + 1 (moving blue by .7 relative to the nation)

Predicted 2020 Winner: Biden by 9 (10 or 11 wouldn't surprise me a bit)

2008: Obama + 12.5 (blue + 5.43 (3rd party 3.2%) (967,848 votes cast)

2012: Obama +6.68 (blue+ 2.78) (3rd party 1.97%) (1,014,918 votes cast)       

2016: Clinton +2.38 (blue + .28) (3rd party 6.5%!!!) (1,125,385 votes cast) 

Final 2016 RCP average GOP + 0.8 Deviation Dem + 3.2    

2018 Senate: Dem + 5 (red +3) (3rd party about .5%) 972,132 votes cast)     

RCP average on 10/7/2020: Biden +5.3 (Red + 3.8)   

Comments on Nevada:

Nevada was/is hit harder than (virtually) any other state by this Covid recession, due to its huge dependence on tourism.  That could hurt Trump here.  Nevada is somewhat hard to poll.  In addition, Democrats typically overperform their polls here.  I have more uncertainty in my Nevada forecast than most other states with the vital exception that I am super confident Biden will win it.

North Carolina (15 electoral votes)

Predicted 2020 lean: red +7 (moving red by 1.25 relative to the nation)

Predicted 2020 Winner: Biden by 1

2008: Obama + 0.32 (red + 6.88) (3rd party .89%) (4.311 m votes cast)

2012: GOP + 2.04 (red + 5.94) (3rd party 1.3%) (4.505 m votes cast) 

2016: Trump +3.66 (red + 5.76) (3rd party 4%!!!) (4.742 m votes cast) 

Final 2016 RCP average GOP +1 Deviation GOP + 2.7    

RCP average on 10/7/2020: Biden + 1.4 (Red + 7.7)   

Comments on North Carolina:

North Carolina has a lot of white southern voters, who have not moved towards Biden anywhere near as much as the overall population.  In fact, these voters may move slightly towards Trump.  North Carolina also has more than average suburban voters, who have moved sharply away from Trump.  North Carolina has slightly more black voters than average, which won’t move towards the democrats, as black voters were about 90-10 for Clinton, but there could be slightly increased turnout.   Since white southern voters are the largest group, and will likely hold strong for Trump, I see North Carolina getting a bit redder.

In the coming several election cycles North Carolina is likely to move a bit towards the democrats (more moderate and liberal voters moving from the north) and is my expected hot new swing state for 2024 and beyond.  By swing state, I mean a state that could go either way in a fairly close election.

North Carolina Senate race: This is perhaps the hottest Senate race this year.  Money has POURED into this race from all over the solar system.  Incumbent GOP Senator Thom Thillis has COVID (get well Tom!) while democratic challenger Cal Cunningham is in a sexting scandal.  I predict (guess, really) that these two things will cancel each other out, and that voters in NC have made up their minds in both the presidential and Senate races.

Republican Thillis is mildly unpopular in this light red state, which should call for a VERY close election.  The 538 model currently predicts the democratic challenger will win about 50-47.  That seems like too big a margin to me, yet I still predict the democrat wins here.  Again, Biden’s 8 point national lead is just too much.

Ohio (18 electoral votes)

Predicted 2020 lean: Red + 7.8 (moving blue by 2.43 relative to the nation

Predicted 2020 Winner: Biden by 0.2.  Ohio could easily go to Trump of course, if Biden wins by 8.  If Biden wins by 9 or more, I’m very confident he’ll win Ohio.

2008: Obama + 4.6 (red + 2.6) (3rd party 2.6%) (5.7 22 m votes cast)

2012: Obama + 3 (red + 0.9) (3rd party 1.63%) (5.581 m votes cast) 

2016: Trump +8.13 (red + 10.23) (3rd party 4.7%!!!) (5.496 m votes cast)

Final 2016 RCP average GOP + 3.5 Deviation GOP + 4.6

2018 Senate: Dem + 6.8 (red +1.2) (3rd party about zero) 4.417 m votes cast)   

RCP average on 10/7/2020: Biden + 1.2 (Red +7.9

Comments on Ohio: Long a crucial swing state, Trump blew Clinton out, winning by 8.  However, Ohio is chock full of white voters without a college degree, and while these voters noticeably favor Trump, he is not performing nearly as well with this group as in 2016.

Given that I predict Biden to win Ohio by only 0.2, I’m obviously not confident in it.  Ohio will not matter in this election, barring lawsuits invalidating large numbers of mail in votes, or other drastically unusual items.

Pennsylvania (20 electoral votes)

Predicted 2020 lean: red+ 1.5 (moving blue by 1.3 relative to the nation))

Predicted 2020 Winner: Biden by 6.5

2008: Obama + 10.32 (blue + 3.12) (3rd party 1.38%) (6.015 m votes cast)

2012: Obama + 5.38 (blue + 1.48) (3rd party 1.43%) (5.754 m votes cast) 

2016: Trump +0.72 (red + 2.82)  (3rd party 4.36!) (6.156 m votes cast) 

Final 2016 RCP average Dem + 1.9 Deviation GOP + 2.6   

2018 Senate: Dem + 13.12 (blue + 5.12) (3rd party 0.62%) 5.009 m votes cast)   

RCP average on 10/7/2020: Biden + 6.3 (red +2.8)

Comments on Pennsylvania:

Pennsylvania has been considered THE most important state of this election cycle.  I agree strongly with this assessment.  If Biden wins by 8 or more, he will easily win the electoral college.  But, hypothetically, if Biden only wins the popular vote by 4, he would NOT win Florida, North Carolina or (most likely) Arizona.  That could make Pennsylvania absolutely crucial, and perhaps decisive. 

Nate Silver has coined the term tipping point state – the state that puts a candidate over 270 votes, meaning s/he has won other states by more than s/he wins the tipping point state by.  I think its VERY likely that Pennsylvania will be the tipping point (as Florida so famously was in 2000, and Ohio was in 2004.  The difference between this election and those 2, obviously, is that I do not expect this election to be close.

Texas (38 electoral votes!!!)

Predicted 2020 lean: Red + 10 (moving blue by 1.1 relative to the nation)

Predicted 2020 Winner: TRUMP + 2

2008: GOP + 11.8 (red + 19) (3rd party 0.97) (8.087 m votes cast)

2012: GOP + 15.8 (red + 19.7) (3rd party 1.45%) (7.994 m votes cast) 

2016: Trump +9 (red + 11.1) (3rd party 4.5!) (8.969 m votes cast) 

Final 2016 RCP average GOP + 11.7 Deviation Dem + 2.7    

2018 Senate: GOP +2.56 (red + 10.56) (3rd party 0.78%) 8.732 m votes cast)    

RCP average on 10/7/2020: Trump + 3.2 (Red + 12.3)  

Comments on Texas: There has been a ton of chatter in democratic circles about flipping Texas to blue in 2020.  That was never in the cards absent a big Biden landslide.  At a guess, he’d need to win the popular vote by 11 or so to flip Texas.  That’s not totally off the table, but I’d sure bet against it.  And more to the point, if Biden wins Texas, he’s won Florida (!), North Carolina, Arizona, and the Midwest, and EASILY won the electoral college.  Texas is getting bluer, and if this continues, could well be in play in 2028 or 2032.  But democrats were dreaming about putting Texas in play in 2020.  I do note that Biden has bought ads in Texas recently.

I see this as a candidate flush with money (as Biden is, another thing I totally did not expect) trying to lure the other side into wasting money, as well as focusing on party building and a series of house seats that absolutely could flip.  Biden’s team is spending only $6 million (a drop in the ocean in wildly expensive Texas), so it’s a what the heck kind of move.  I would bitterly oppose spending real money.  I think Team Biden would be much better served to spend additional money in Iowa and Ohio (they are, I assume, maxed out in North Carolina, Pennsylvania, Wisconsin and Arizona) and as for Florida, Mike Bloomberg has promised to spend $100 million to help Biden.

Virginia (13 electoral votes)  

Predicted 2020 lean: Blue + 5.5 (moving 2 points towards blue relative to the nation)

Predicted 2020 Winner: Biden by 13.5

2008: Obama + 6.3 (red + .9) (3rd party .9%) (3.723 m votes cast) 

2012: Obama + 3.9 (blue +0) (3rd party 1.5%) (3.855 m votes cast)

2016: Clinton + 5.6 (blue + 3.5) (3rd party 5.9%) (3.985 m votes cast) 

Final 2016 RCP average Dem + 5.3 NO DEVIATION

RCP average on 10/7/2020: Biden +11 (Blue + 1.9)

Comments on Virginia: Virginia has, since 2000, gone from light red to purple to light blue to flat out blue.  The large number of people dependent at least partially on federal government work has put Virginia out of reach for the GOP absent a landslide win.  I’m not saying the GOP can’t win the popular vote by 5.  I think it could.  But unless it does (and perhaps by even more than 5) Virginia is now safely in the blue column for a long time.

Wisconsin (10 electoral votes)

Predicted 2020 lean: Red + 1 (moving BLUE by 1.88)

Predicted 2020 Winner: Biden by 7

2008: Obama + 13.9 (blue + 6.8) (3rd party 1.47) (2.983 m votes cast)

2012: Obama + 6.94 (blue + 3.04) (3rd party 1.45%) (3.068 m votes cast)

2016: Trump +0.77 (red + 2.88) (3rd party 6.33!!!!!) (2.976 m votes cast) 

Final 2016 RCP average Dem + 6.5 Deviation GOP + 7.2

2018 Senate: Dem +10.83 (blue + 2.83) (3rd party 0.1%) 2.661 m votes cast)     

RCP average on 10/7/2020: Biden + 5.6 (Red + 3.7)    

Comments on Wisconsin: Earlier on in the cycle everyone including me expected Wisconsin to be reasonably close.  The fact that Trump has lost significant ground among white voters without a college degree is fatal in Wisconsin.  He won it by less than 1 point last time.  He has no margin for error, and the move among these voters has been significant. 


Sunday, January 03, 2016

I predict the S&P 500 will end 2016 at 2,336, up about 14% from its year end 2015 figure of 2,044

Cliff notes

I predict that the S&P 500 will end 2016 at 2,336.  This represents a 14.3% increase above the year end 2015 figure of 2,044.  Barring a black swan event, I think the risks to this forecast are slightly to the upside; that is I think we could see the S&P above 2,400.  Note that the current all time high for the S&P 500 is 2,134, so I am predicting a series of fresh all time highs.

I would likely predict about the same 14% increase for the Dow Jones.  It closed 2015 at 17,425, so I would predict it ends 2015 at about 19,865.  Obviously 20,000 would be very much achievable if my forecast is right.

I note that I have been bullish on the market for a LONG time, and remain so.  More on that in an upcoming blog post.

I arrive at my prediction of 2,336 for the S&P 500 by assuming about a 14% increase in earnings for the 500 ish companies in the index, while keeping the price to earnings multiple the same. .

I expect the US economy to grow about 2.5%, just slightly above the very boring 2%+ range it has been in for much of the recovery.  (I must note here that I have been noticeably too optimistic about US GDP growth for years).  I expect interest rates to rise gently (perhaps 50 basis points on the 10-year treasury bond).  I expect inflation to pick up once oil bottoms, but not in a fashion that will scare the Fed or the market.  I expect the fed to raise interest rates only 2 or 3 times in 2016, as a still sluggish economy, and slightly slowing job growth mean that a real inflation scare is very unlikely.

I arrive at my 2016 Price Target by Estimating S&P 500 earnings and estimating a Price to Earnings Multiple

One way to value a stock or a market is by the age-old price to earnings ratio.  If a stock trades at $10/share, and earns $1/share, it has a Price to Earnings ratio (P/E) of 10 (and is often considered cheap).  The same valuation methods work for the S&P 500.  However, this is not a simple process.  First, you need to predict the earnings, or the E.  This is tricky enough, and estimates 12 months ahead are often well off the mark (typically too optimistic).  Next comes the harder part. You must assign a multiple.  Simplistically, if a market (or a stock) trades at a low multiple, say 10, that means the market is pessimistic on future earnings growth for that company.  Interest rates also play a large (and in my opinion hugely under appreciated) part in assigning an appropriate multiple.  Low interest rates should and typically do mean high P/E multiples, and high interest rates mean low multiples.

Thus in order to arrive at a price target for the S&P 500 by this methodology, I need to (a) estimate earnings for the S&P 500 in 2016; (b) estimate the Price to Earnings multiple at year end 2016; and (c) multiply the earnings by the multiple.  The result of that multiplication is my price target.  Is this a valid exercise?  Yes, I think so.  There are a series of assumptions that go into my estimate of the "appropriate" multiple, and disagreement with these can easily yield strong disagreement on the multiple I assign, and a wildly different price target.  I don't mean to make this sound scientific.  But I humbly submit there is some science involved.

S&P EPS 2015

Earnings per share for the S&P 500 for 2015 were about $105/share.  That is the number I am using.  (Its impossible to know exactly as final results are not in yet and I found significantly different numbers in a few places, some lower, some much higher.  

Raymond James said just a few weeks ago that 2015 earnings were $107.   

Standard & Poors' spreadsheet, with some extrapolation, yielded $105, so that's what I'm going with.   (additional Info, Index earnings).

I note that as for the obvious question what does S&P 500 earnings per share mean, the issue is quite complex. There is no one company, obviously. Instead, its a formula, based on the earnings of the constituents of the S&P 500, multiplied by the company's weight in the index.  Thus Apple's earnings count for more than General Electric, as Apple's market capitalization is nearly double GE's.  And so on, with smaller companies counting for less.  Honestly, I don't really understand the formula, except in broad outline.  Happily, I don't have to.  By relying upon published actual and estimated earnings per share and using the closing price of the index, I can easily compute the price to earnings multiple (as many do) and go from there.  I don't really need to understand precisely how it is computed.  

S&P 500 Price to Earnings Ratio for 2015.

The S&P ended 2015 at 2044 (rounded to the nearest whole point).  Thus the Price to Earnings ratio at year end 2015 was 19.47 (2044/105 = 19.47).  I use this P/E ratio to come up with my 2016 target for the S&P 500.  I assume no change in the multiple at all.  While this is obviously very unlikely to occur precisely, it does mean that I am not relying on any expansion of exuberance on Wall Street.  

S&P P/E recent history

19.47 is a fairly high P/E ratio historically, but not wildly so.  the ratio was MUCH higher during the wild bull market of the late 1990s, when I think nearly everyone agrees the market was significantly overvalued.

I think 19.47 is slightly below where the P/E ratio should be, given that inflation is VERY low, interest rates on long term government treasuries (key competition for bonds) were VERY low compared to history, and earnings growth potential for the next few years was/is good.  Obviously, one can disagree with these key assumptions.

One way to look at this is what is called the earnings yield.  This is merely the inverse of the Price to earnings ratio.  If the P/E ratio is 20 (close enough to the actual 19.47 for this big picture point) that meanings the Earnings yield of the S&P 500 (earnings/price) is 5%.  Put another way, stocks are "earning" 5%.  $105/share in earnings is just under 5% of the 2,044 that the S&P ended the year.  That's a pretty good deal in comparison with the 10 year treasury bond, which closed the year at 2.27%.

Predictions for S&P 2016 Earnings Per Share 

I am "estimating" 2016 earnings to be $120/share, a number I believe to be slightly conservative.  When I say I am estimating, what I really mean is that that is the number I am ASSUMING.  I lack the capacity (by a long shot) to actually run bottom up estimates for what each of the 500 (ish) companies in the S&P 500 index will earn and then multiply each appropriately.  Instead, I am relying upon estimates by the big banks.

RBC (Royal Bank of Canada) predicts S&P earnings for 2016 of $124/share (recently cut from an earlier $128).

Raymond James predicts $126/share.

Goldman Sachs in late September predicted $120/share for 2016:

Barrons states that the mean average Wall Street estimates for 2016 are $123.50, with most estimates being higher.

I am assuming $120/share.  $120 represents about a 14% increase in earnings over 2015.  A 14% increase in earnings may seem very optimistic, but is in line with earnings estimates by Wall Street, above, (which I freely admit are almost always too high).  I think this time the estimates are slightly too LOW because I expect the consumer to finally spend more of the savings resulting from massively lower oil prices, while the negative effects of lower oil should be largely played out.  If it sounds like I am saying this time is different, that's because I am.

By coincidence, the $120 estimate I am using for 2016 is exactly the same as Goldman's estimate.  I think the risk is to the upside; that we could actually end up between 125 & 130, as the more optimistic seers on Wall Street are predicting.  However, I have been too optimistic on economic growth (which tends to yield higher earnings) for a long time now, and its time to make a different mistake!

S&P predicted 2016 P/E and thus target price for 12/31/16

2016 earnings per share of $120 X the projected P/E ratio of 19.47 = my target of 2,336.

The final item necessary for me to come up with a price target for year end 2016 is to estimate the Price to Earnings multiple at the end of 2016.  This depends on my view of interest rates (gentle rise), inflation (rise), and my prediction for overall optimism of the market in general (about where it is now, or slightly more optimistic).

As I said above, I am using the P/E ratio from year end 2015, which was 19.47.  I think that forces pulling it higher are evenly balanced with forces pulling it lower.

Forces pulling the P/E higher include slightly better economic growth, a predicted encouraging rebound in earnings, better numbers out of Europe and possibly Japan, China failing to implode and slightly less pessimism about the US recovery.

Forces pulling the P/E lower include slightly higher interest rates, continued sluggish growth, potentially a further rise in the dollar (this would surely effect earnings, whereas the effect on the P/E multiple is not clear), and, most importantly, a rise in inflation, as we lap the sharp declines in oil prices.

Crucially, I am betting that Janet Yellen does not panic in the event inflation rises slightly as she and others predict, and raise short term interest rates quickly.  That is absolutely mission critical to everything written here.  If the economy performs as I expect, and Yellen raises rates 5 or 6 times, the market should be about flat for 2016 even if we do earn the $120/share on the S&P 500 that I am assuming.  The reason is that the market will take down the P/E ratio in a world of rapidly rising short term interest rates and a sluggish economy.  In contrast, if we see 3%-3.25% GDP growth, the market would happily accept 5 or 6 interest rate hikes.  This is a very unlikely scenario.

A final word about inflation.  Inflation should FINALLY move towards or even above the Fed target of 2%.  With the recent sharp price DECREASES coming off the books, and a predicted somewhat higher oil price for the year, the inflation in the rest of the economy (health care, education, rent, housing) becomes a larger weight in the CPI. See Bob Johnson of Morningstar.