POLITICS

INFO

Possibly of interest to other Wisconsinites. On the site My Vote Wisconsin you can get both past and present information concerning your voting. This includes information about current absentee voting for the 2020 presidential election.

Mine shows that my absentee ballot should of arrived around Sept. 17th, which it did. It now shows that my returned ballot was received yesterday Sept. 22.

I VOTED

My thanks to the Municipal clerk for the Town of Hudson, St. Croix County, and all involved for permitting a fair and honest election.

So far these all involve using the R language and RStudio to do some data analyis related to politics. This is also an interest of mine. A skill I am attempting to work on a little bit although I am no expert.

The R source and datafiles are on github. Links will be provided for each project.

Election 2024

Electoral College vs. Popular Vote 2+ Revisited Again

Ya can't tell the players without a scorecard

I was watching the news and there was a Trump rally in North Carolina. I wondered, what could what I had done tell you about North Carolina and why Trump might be there. It shows that in 2000 North Carolina wasn't a swing state. In 2016 it was 10th (closest) in swing states and went Republican with them showing a greater distance gain in turnout. In 2020 it was 5th (closest) of the swing states. It again went Republican but showed a greater turnout gain for the Democrats. Explaining it's move up the swing state list, it got closer. With 16 Electoral College representatives (2024, 2020 shows 15), it would be a state Trump would want to hang onto. With it's closeness and seeming Democratic trend it could be a state the Harris campaign would want to 'put into play'.

I thought more content on interpretation of the data could be useful. That it might be handy as a reference for election related as things progress.

Interpretation

One thing I thought was that maybe the data on 'distances' wouldn't immediately make sense. So I added the actual turnout percentages in the swing state section. You can maybe see how close the very first swing states are. Usually not more than a tenth or two of a percent separation.

The PvEC_2 Swing state related...
Swing States And Turnout

It seemed that it might be interesting to know if these differences were consistent election to election. I broke each of the three swing state section elections into three 'closeness' ranges, 1:5, 6:10, 11:15. Originally I did this only for the two states of interest for PvEC, 2000 & 2016 plus I also added 2020 as the latest and I thought more typical election. Looking at that shows that the 2020 numbers were larger.

Sometimes in doing this I have wondered if it doesn't just always boil down to number of votes. The more votes the bigger the numbers for whatever you look at. That could be the case here since the 2000 and 2016 were clearly closer.

Turnout and Polarization

Although, it could also mean that the 2020 results were more polarized. Possibly, this could be determined if you looked at more elections. I did this and got the following, reproduced from the link above.

Swing State Turnout Differences In Ranges

##       Year        1_5      6_10     11_15
## tdiff 2000  0.1311604  1.590966  2.680234
## X     2004  0.6980677  2.082592  3.697898
## X.1   2008 10.7027048 16.885724 11.887326
## X.2   2012  9.4495818 14.396069 10.566800
## X.3   2016  0.4241536  1.757940  3.915853
## X.4   2020  0.4987496  2.775959  6.052002

Returning to the original question of whether or not the Biden-Trump election showed polarization. Maybe. Certainly not on the scale of the Obama elections, but as the states get less close there are larger differences in turnout, which would be electorate polarization. This would mean one or the other party is outperforming for turnout in most cases. Which in turn would make swinging or flipping the state that much less likely.

Swing States and Campaign Appearances

It has seemed to me, and this is an impression, that the candidates for this election have been concentrating on very few states for appearances. The closest of the swing states. Wisconsin, as I've mentioned, we rock the vote for turnout but are also consistently a very close swing state. Refer to the link above, we were top 5 in all three elections, 3rd in 2000, 4th in 2016, and 3rd again in 2020. Then Pennsylvania, 3rd in 2016, 4th in 2020. Finally, the appearances in North Carolina, 10th in 2016, moved to 5th in 2020. These are the states I've mostly been aware of, just following the news a little bit, where the candidates have been making appearances. I have checked a little for mention of appearances in other places but didn't really find any.

There could be various reasons for limiting appearances to the closest of the 2020 swing states. Given Harris assuming the top spot in the Democrat ticket there has been talk that the remainder of the election is a 'sprint'. Maybe the campaigns are trying to get a jump in these states and will spread out later. Maybe beyond the top 5 closest the rest are in fact considered too polarized to flip. Maybe with nation wide media appearances are considered less important. There is also the internet, Trump with social media and my understanding is Harris has some content and support online. Maybe these states have been calculated to be the "key battleground states" that the election will hinge on.

One thing I have found sort of surprising is that no one that I've heard about has been in Georgia or Arizona. The swing state section shows these as 1 and 2 for closest in 2020. Biden took both but it required a considerable gain in turnout. It definitely doesn't seem a lock for 2024. I did hear the Democrats were spending money on ads in these states despite them not being "blue wall". Neither is North Carolina. I remember one of the 2016 post-election criticisms claimed was that Hilary Clinton didn't spend enough time in the states that she should of. I wouldn't take these states for granted on either side.

I did just hear that Robert Kennedy had appeared in Arizona to sort of weirdly partially endorse Trump, so Trump has been there. Also this morning I heard Harris is going to be in Georgia, so these states are not being neglected.

As I watched the DNC convention the states most often getting mentions include these two and the others from the closest 2020 swing states as shown in the prior link. So the swing ranking numbers seem to match up with the awareness of their importance at the convention.

The Close And Electoral College Elections

I heard it said that in 2016 Hilary Clinton lost because she failed to excite the voters. This seemed to me like accusing the victim. The numbers do provide some confirmation. If you look at the 2016 election here many of the Democrat turnout results were below average and in the top 10 the Republicans had better turnout in every election. I tend to think of this more as a letdown from the interest in the Obama elections rather than a fault in Clinton. I think this would have happened with any follow-up candidate.

Choosing not to exercise your right to vote is a valid choice in this democracy. But, don't come back later and say the dog ate your vote. I think if something unexpected happens people want to come up with an explanation, or if its not good, an excuse.

If you look at the other two elections shown you see that in 2000 it shows turnout down for both parties. Again, though the data isn't adjusted for population increases and 2000 would be at the far low end. The data table shows that 2000 had the 2nd lowest "United States" turnout at 50%. The summary section might again give an idea of the 'closeness'. 2020 was more a mixed bag all positive, record turnout, some more Democrat some more Republican. The Democrats had slight wins in the swing states in both 2000 and 2020 while losing big in 2016.

I'm not going to say you can look at the turnout distances I have and say what the outcome probably was. I couldn't look at the 2016 numbers and say it seems Clinton lost pretty big in the swing states. Some states show a party improving in turnout more than the other while still losing the state. It is meant to provide a 'relative' estimate of how turnout went. Did the state improve over average or get worse? Did the party improve or lose turnout more than the other party? Can you tell to some limited extend if there is a trend. A state becoming more or less swing. Anyhow, some information you don't get strictly from number of votes for the year.

In 2020 the Republicans turned out in force. This time, however, the Democrats responded in kind. It was a record turnout year. Given this, I believe the advantage that the Democrats normally have, even with the Electoral College, handed them the victory.

Gerrymandering

My original idea concerning Electoral College only wins being caused by something like gerrymandering wasn't correct. However I think the metric I tried - Gap Per Elector(GPE), wasn't completely off the mark. From the Gap Per Electoral College Representative(GPE) part of the link. Considering again both Bush elections and the Clinton 2016 election. These show as the closest elections by both popular vote and GPE. I still find thet 2004 election of more interest here than the 2000. Wikipedia shows that Bush won by 2.4% of the popular vote. In 2016 Hilary won popular vote by 2.1%. Bush won a clear electoral college victory while Clinton lost. Again sort of suggesting the casino nature of the Electoral College. GPE is actually closer, less extra votes in 2008 than in 2016. Also note that both distance measures agree on direction, positive for a Democrat year, negative for a Republican. Except for 2016. GPE does show it as negative year, more gap or votes to spare for Trump. I think somewhat as a surprise to many that he had found a constituency.

The GPE sort of combines information from popular vote and the electoral college. I think there is unique information that can be interpreted from that which isn't just how many votes.

As an after thought. When I was thinking the electoral college sort of represented a gerrymandering situation for presidential elections. Since determined to be incorrect. Still, I thought, a possible solution to make the Electoral Collega a fair system would be the same solution for states to avoid gerrymandering. Redistrict it. Of course you couldn't just change state borders to manage this. But you could probably come up with something nation wide at the county level where you could get regional districts more equitable than by states. Although, "Electoral District 72, Electoral District 72, Electoral District 72" wouldn't have the same ring to it as "Florida, Florida, Florida".

That is farfetched and no one would go for it. Even though it could be a fix to make the Electoral College system more fair. As it is I still believe that the Electoral College system gives the Republicans a builtin edge in close presidential elections. I also still think there would be an almost insurmountable edge for the Democrats if the elections were strictly popular vote decides all. The Republicans would rarely if ever win the presidency on this basis. Both my simulations, Previous simulations also including 2020, and historical results since 1992. The Historical is shown in the summary from the current update of what I had originally looked at for this.

The simulation results above show an 88% chance the Democrats win on popular vote. The prior results which didn't include 2020 were actually Democrats winning 90.81%. The simulation results are based on some randomness. Adding another Democrat popular vote win should of made it more likely probabilistically speaking. The simulations also didn't include the Bill Clinton Democrat popular vote wins in 1992 and 1996 which should of made Democrat wins this way more likely. So it's sort of safe to say going by the simulations the Democrats should win more than 9 out of 10 presidential elections.

Historical, the only time in the years considered that the Republicans won on popular was the George W. Bush re-election. And that was a close thing as mentioned earlier. So an actual 1 out of 8.

I don't believe the Electoral College should not be abolished in favor of popular vote. I believe in multi-party democracies. Minimum two parties. There should be representation for opposing viewpoints. I think voter suppression is absolutely undemocratic and unamerican. I think the Republicans totally gamed the Supreme Court justice selection process making the rules whatever gave them an advantage. Which was hypocrisy. In turn, giving us possibly the most partisan politcal court ever. But, Republicans still represent the views and values of a lot of people and I don't think they should be completely shut out of chances for the presidency. I believe popular vote only elections would just about amount to that. If they never win the presidency we are in effect no he longer a two party democracy, although there is more to the government than the presidency.

In Closing

So, to close with some entertaining type references. Stevie Wonder appeared at the Democratic convention. I remembered the first place I had seem him was al little Stevie Wonder in the old beach movies.
MUSCLE BEACH PARTY (1963) - (MOVIE CLIP) LITTLE STEVIE WONDER>
I remember now that they always had go-go dancers in these scenes. Hopefully happy street is your street after the next election.

This one I hope won't be taken the wrong way. But the current Democrat theme of "We're not going back" reminded me of this one.
Twirling, Twirling, Twirling Towards Freedom! (The Simpsons)
Really, if it suits you to do so, feel free to twirl your way to freedom.

I thought I had some insight concerning the Electoral College only wins connected to Gerrymandering. This turned out not to be the case. The results are similar to gerrymandering in result only, not in mechanism. There are possibly some things of interest in the derived metric - Gap Per Elector (gpe). Mainly using actual historical results I consider turnout and again swing states.
New Content here...

Popular Vote vs. Electoral College 2024 - Part II

Electoral College vs. Popular Vote Revisited

I was originally going to do a rather simple update of my presidential election simulations, just adding the 2020 results. I ended up doing a little more than that. It still seems that presidential elections resulting in a Electoral College victory while trailing in popular vote are not real unlikely and they are more likely than previously. They are still not the most likely outcome though.
A new look at that is here...

Popular Vote vs. Electoral College 2024

I have left the old one...
Popular Vote vs. Electoral College

It occurred to me in doing this that what this situation amounts to, where the win is in the Electoral College while losing popular vote. This is the same as Gerrymandering. There are obviously 'wasted' Democrat votes. If you win by one vote that is 'efficient'. Winning by more that, the votes are wasted and it's inefficient.

This has been called an Partisan Gerrymandering and the Efficiency Gap. It is considered a sort of measure of the extent of Gerrymandering. It has been used in legal cases to decide if states have been districted fairly. I have looked at it some myself Gerrymandering below. Or, the web page generated from my R coding is at Gerrymandering.

I am not saying anyone has purposely laid the states out to win presidential elections. There is no need for an investigation or law suit. But I think just how things have turned out demographically and geographically the Electoral College sometimes acts as Gerrymandered for presidential elections. Using some of the data from above I would like to calculate efficiency gaps to see to what extent this is true. Yet again, it seems like it should be fairly simple.

I hadn't meant to upload anything until the Gerrymandering part was also done. But since the Electoral College v. Popular is pretty much independent I am posting it first. Consider it like a leaked Supreme Court decision at this point.

Election 2020

Electoral College vs. Popular Vote

There are interesting things starting to come out of the 2020 election besides the Mueller Report, very much in the news today, and a platoon of Democrat candidates. One thing that I just heard which got me thinking was Elizabeth Warren making the elimination of the Electoral College in favor of popular vote a campaign issue in her run for president. I got to thinking this might not really be that good an idea. This page Popular Vote vs. Electoral College did sort of turn out to be a defense of the Electoral College as it is.

You might want to take a look if you're thinking 'the Electoral College is so 18th century, let's just go popular'.

A MOOC course I took required signing up on Twitter. I don't use it, but I get emails with potential items of interest from some recommender system. I glance at the emails occasionally and coincidentally noticed this one in the last.
David Weakliem on the U.S. electoral college
So there is what I would have to call topical interest.

It's not like the subject is new though. From the United States Electoral College Wikipedia.

Polls since 1967 have shown that a majority of Americans favor the president and vice president being elected by the nationwide popular vote, instead of by the Electoral College.

The github project is Election-2020

Electoral College Upsets

I was watching the Reidout show on MSNBC and Steve Kornacki was as usual at the graphics board explaining how some polling was currently much more favorable for Biden than it had been for Clinton in 2016. After he finished Joy Reid said "I have one question". I think she wanted to know if this indicated a definite win for Biden. I couldn't find the video clip or transcript so hopefully I'm getting her meaning more or less correctly. I assume she was concerned because of the 2016 election results. Either how the polling had proven incorrect or by the Trump upset victory in the Electoral College.

The concerns are understandable, consider this from the 2016 election Donald Trump Has A 20 Percent Chance Of Becoming President from FiveThirtyEight. That right there was the conventional wisdom at the time. Trump was given pretty narrow chances of winning. But as Nate Silver notes...

How do you predict a general election with Donald Trump?

We can think of a few basic approaches. One of them is to assert that precedent doesn’t apply to this election and that Trump’s case is sui generis. It’s not clear where that leads you, however.

First, you might wonder if the polls can be trusted. I've noticed that these polls generally indicate a margin of error around 3 to 4 percent. So, if Biden is leading by more than that in the polls he should win. However, polls can be wrong. Why experts are getting presidential election polls wrong – again. These polls are based on well founded and well understood statistics for things like proper sample sizes and where errors might come from. But, they can still be wrong, especially it seems for Trump.

I was more interested in the possibility of another Electoral College win contrary to popular vote. Specifically, how much of a popular vote lead would Biden need to be safe from an Electoral College upset? I thought what I did above for election simulations could provide a guesstimate answer to this. I took the same approach, running 10,000 random election simulations based on recent presidential election results. Sort of a Monte Carlo approach. Except this time I just paid attention to the Electoral College upsets, when one side wins in the Electoral College while trailing in popular vote.

The average population difference is what the vote difference usually is when a electoral college upset occurs. So, at 1,851,512, it means that upsets tend to happen with a vote difference of less than 2 million. The max values are our guesstimates for what point the chance of an upset pretty much goes to 0. So, for the democrats if they are losing in popular vote by more than 3,417,713 they have pretty much no chance of getting an electoral college upset win. However, for the republicans they can be behind by up to 7,279,770 votes and still win in the electoral college. I have had runs where that was up to about 9 million votes. In this regard things clearly favor the republicans.

Next, I include the actual number of Electoral College upsets for each party. This is followed by the probabilities for a upset and then upset probabilities for each party. This usually works out as shown to about a 20% chance of a Electoral College upset win, with about a 4% chance of a democrat win this way but about a 16% chance for a republican win. So again, advantage republicans. I didn't notice that these could happen quite so often when I did the original simulations. This is not what I would call a slim chance but seems to give republicans a fairly substantial edge.

Finally, for the Clinton popular vote margin of 2,868,686 I calculate the probability that she should of lost in a Electoral College upset. This usually comes up as shown between 4 and 5 percent. A fairly low chance, unless maybe Trump is involved.

A histogram for the simulations is included

Upset Elections