Tuesday, November 29, 2016

Orange County Vote

People have been commenting on how poorly Donald Trump did in Orange County, California and how that might lead to the downfall of Orange County Republicans. We don’t have Trump congressional numbers yet, but we do know that Mimi Walters, CA-45, and Ed Royce, CA-39, did roughly the same as John Campbell and Royce did in the 2012 election. Dana Rohrabacher, CA-48, dropped 3 points. Darrell Issa, CA-49, was the only Orange County Republican with major problems. His problem stemmed more from the San Diego portion of the district. He won that by roughly 22,000 votes in 2012 but lost it by roughly 14,000 votes this year. That's a huge swing.

There wasn’t one Democrat and one Republican on every congressional ballot but there was on the assembly ballot. There were roughly the same number of two party Presidential votes as there were assembly votes. The voter drop-off roughly corresponded with the third party Presidential vote. Trump lost 54.6%=45.4%, but Republican assembly candidates won 56.6%-43.4%. Trump got 492,152 votes. GOP assembly candidates got 615,755.

Darrell Issa clearly has a target on his back in every election from here on. The other three Republicans all managed to get at least 57% of the vote. There's no indication they'll be in trouble for at least the next two cycles. Redistricting will take place for the 2022 cycle. I have no idea how that'll impact the districts. Replace a Republican city or two from Ed Royce's district with Democratic ones and he'll lose too. Fortunately for Republicans the lines won't be drawn by Democrats but by an independent commission.

Sunday, November 27, 2016

OC Registrar of Voters Responds Inaccurately

The Orange County Registrar of Voters has responded to suggestions of irregularities. It's full of inaccurate information and addresses the wrong issue. They assume the issue is turnout percentage but I think Jim Brulte is talking about the number of voters. Currently nationwide turnout is 4.3% higher than 2012, California turnout is 1.9% higher than 2012, and Orange County turnout is 5.0% higher. These numbers should grow some as more ballots are counted but nationwide won't hit 6% higher and California will probably be 7-8% higher. If all remaining Orange County ballots are valid they'll be up 9%. Of course SD-29 will go up too.

As of now SD-29 currently has 10% more ballots than 2012, twice the percentage of the county as a whole, and five times the percentage of the state. I assume part of the reason the county is higher than the state is because of the additional ballots in SD-29. I have no idea what the numbers are precinct by precinct by Jim Brulte is reporting that the number of ballots is up 30% in some precincts.

I'm not suggesting there's fraud, but it'd be really odd if there were precincts with 30% more votes in a state with 2% more votes. And really I just prefer people use accurate data.

Saturday, November 26, 2016

CA GOP Chairman Jim Brulte Thinks Something Funny Is Going On

Jim Brulte, the California GOP chairman, sees something suspicious with unusually high turnout somewhere in Orange County. Since Brulte is only going to be concerned in a district where Republicans stand to lose a seat, it's likely that he's talking about SD-29. This is Bob Huff's district right now, but he's termed out. Assemblywoman Ling Ling Chang led the race by 3,923 at the end of election night. On November 16 that lead was 5,051. When additional ballots were counted on Friday, November 18, Chang's lead dropped to 187. Democrat Josh Newman then took the lead in the next count by 829 votes and is now up by 1,650.

I've been tracking the counts in 55 congressional, senate, and assembly districts and Democrats have done better in 39 of them. While Democrats often do better in post-election night counts, it's not universal. Newman had 49.1% of the vote on election night and has gotten 52.8% since then. That shift is the 7th largest in favor of the Democrats in the post-election counting. In the other cases, however, the counting has consistently been in favor of one party or the other. A reversal like this one if fairly unique.

Orange County doesn't list turnout by precinct until the results are certified. The Orange County portion of SD-29 is up by 10%. This might come heavily from some precincts which are up 30%. The smallest Orange County does go is by AD. Here are the AD vote increases:

AD-55: +7%
AD-65: +7%
AD-68: +2%
AD-69: +19%
AD-72: +15%
AD-73: even
AD-74: -2%

This turnout does look a bit odd, but I don't know if such a disparity is unusual. It's work noting that AD-69 and 72 have the largest turnout increases and I don't think any portion of either is in SD-29. I think all of the Orange County portion of SD-29 is in AD-55 and 65. While SD-29 is as large as two assembly districts, some of the state senate district is in San Bernardino and Los Angeles counties. So the portions of SD-55 and 65 that are in SD-29 might be up more than the portions that aren't. Brulte likely has precinct level data and, I assume, he can tell if there were certain precincts that were giving Newman 48% before November 18 and 58% after it.

There's intrigue here but until there's more information all we can do is speculate.

Wednesday, November 23, 2016

No, the Results Weren't Hacked

The Internet is abuzz due to an article in New York Magazine that alleges that there's something fishy in the Wisconsin tally because "Clinton received 7 percent fewer votes in counties that relied on electronic-voting machines compared with counties that used optical scanners and paper ballots. All they've done here is point out what they feel is an anomaly. The people at 538 debunk the theory when controlling demographics. The people with this theory offer no theory of how these results were fixed. New York decides the voting machines were hacked, even though there's no evidence of hacking. Hacking is often detected. When the DNC was hacked, they discovered it. Of course it's possible people hack and we don't detect it. If that happens, what's the point of asserting it? We can't prove it.

Let's say there was hacking. How would you do it? Well, the problem with hacking an election is that the ballots aren't counted in one central location. There are 3,620 precincts in Wisconsin and each counts their votes and then transmits them to the state. If the vote totals were hacked at the state level and didn't match what the precincts transmitted, people would know. So you'd have to hack the precincts. Wisconsin was decided by roughly 26,000 votes. The final Wisconsin polls showed Hillary Clinton winning by roughly 6%. To assure Trump would win you'd want to manipulate over 100,000 votes. If you hacked into 500 precincts you'd have to change you could change each precinct by taking 200 votes from Clinton and giving them to Trump. That sounds like something that might not arouse suspicion, but that's a lot of voting machines to hack and not be detected. And that assumes the polling was accurate. What if Clinton actually beat the polling the way Barack Obama did in 2012?

Flipping Wisconsin back to Clinton wouldn't change the electoral college vote. Trump would still win. You'd have to also flip Pennsylvania and Michigan. Trump was polling closer Colorado, New Hampshire, and Nevada than those three and was supposed to do better in Virginia than any of those. If you didn't want to be detected and wanted to arouse the least suspicion I'd think you'd go after those four states. Of course it's possible they did and they failed. So your number of states they would've had to hack would be anywhere from 3-7. Did the Russians hack into 3-7 states and no one detected it? We certainly won't find out by doing a recount. It is undetectable after all.

Some people on the left are saying that since Trump had narrow victories in these 3 states where the polls showed Clinton ahead by 4-6 points there must be something fishy. California polls showed Clinton winning by 20-26%. Clinton is winning the state by 31.5%. If we use the theory that beating the polls means there’s fraud, then there’s some massive fraud in California. Clinton stole 350k votes from Trump! Trump fixed states that helped him win. Clinton fixed a state that didn't help her at all. Or something like that.

In Defense of the Electoral College

There's been an outcry, mostly from Hillary Clinton supporters, that the President should be the person who gets the most popular votes. Hillary Clinton is nearly 2 million votes ahead and her lead keeps getting larger. Yet she won't be President. And it's not a travesty.

Defenders of the national popular vote winner claim that America is the only nation with an electoral college. The term "electoral college" is likely uniquely American, so sure, we're the only ones with one. The U.S. isn't the only nation that doesn't determine its head of state by who got the most popular votes. Many nations have a parliamentary system. Their prime minister is determined by an election in one of their houses of government. The people who elect the prime minister are elected by voters in regional districts. This'd be roughly the equivalent of the Speaker of the House being the head of the government in the U.S. In 2012, Democrats won the House of Representatives vote and yet the Republicans had the majority in the House. If the U.S. were the U.K. Paul Ryan would've been President. In 2016, Republicans won the House of Representatives vote. If the U.S. were the U.K. Paul Ryan would be President. The electoral college is pretty much an extension of using voting districts to determine who gets elected. And Donald Trump did win a majority of the House of Representatives districts.

The founding fathers didn't want the winner of the national popular vote to with the Presidency. For the first several Presidential elections most states didn't even have a popular vote. The first election where every state had their citizens vote for President was in 1864. Electors have only been required to adhere to the popular vote in their state in the years since. I'm sure there were delegate to the 1787 Constitutional Convention who wanted the people to choose a President but that's not how the Constitution is written.

This election crystalized why we need the electoral college. President Obama won the median state by 3.0%. Hillary Clinton lost the median state by 3.6%. Donald Trump had a better margin than Mitt Romney in 39 states, while Clinton had a better margin than Obama in 11 states and D.C. The median margin change was 4.2%. Donald Trump did better than Mitt Romney in most places and in over half his margin was bigger than Barack Obama's overall margin of victory.

Hillary Clinton improved on Barack Obama's margin in three big states, California, Massachusetts, and Texas. That's obscured all the places Donald Trump improved due to their sheer size. One thing these three states have in common is that neither candidate was trying to win their votes. Neither campaigned in any of them, spent any money, or mobilized volunteers. Trump did better in the states where the two were actually trying to win votes.

What this shows to me is that if we elected the President by national popular vote the candidates would concentrate their efforts in a few urban areas to the detriment of everyone else. With the electoral system candidates need to compete in diverse group of states that have different needs. Issues that matter in the Rust Belt might not matter in the Sun Belt. Yet a candidate needs to appeal to both. If a candidate concentrates solely on the big cities in a few states there's not much of diversity of interests. Sure, there are some states that don't matter in the current system because they are in the bag for one of the parties but at least more states matter than would matter with a national popular vote.

Tuesday, November 22, 2016

Donald Trump Did Better With Minorities than Mitt Romney

This year's election result has shocked a lot of people. Donald Trump was supposed to have no chance. My friends on the left were predicting a Hillary Clinton 20 point win. Even my friends on the right were saying Clinton would win by double digits. Hillary Clinton wondered why she wasn't ahead by 50 points.

And yet Donald Trump won the electoral college and came pretty close in the national popular vote. One of the things that was supposed to sink Trump was that since he was prejudiced against Blacks, Muslims, and Hispanics not only would he get few of their votes but they'd turn out heavily. While Trump had an appeal to white working class voters, he was supposed to crater with white suburban college graduates too.

Yet he didn't. Trump did a bit better than Mitt Romney among Blacks and Hispanics and actually doubled Romney's vote percentage with Muslims.

Wait. What?

How did this happen?

There were a few reasons. First, there was Hillary Clinton. She wasn't the nation's first Black President and that chipped away just a little at her Black support. She just didn't generate the excitement Barack Obama did and her character issues likely weighed her down across the board.

Then there was Trump. Democrats had painted Republicans as hating Blacks, Muslims, and Hispanics in the past. Everyone who wasn't going to vote Trump because he allegedly hated these groups were already firmly in the Democratic camp. On the other hand, there were minorities voting Obama who didn't base their vote on their belief that Republicans hate them. If they didn't vote for Obama because Mitt Romney hated them, they weren't going to vote for Hillary Clinton solely because Trump hated them. Many of them were blue collar. Trump had a message that spoke to blue collar America. Many people thought this message was reserved for blue collar whites, but it certainly resonated with other groups. Getting 13% of Muslims isn't a good performance, but it shows that 5% of Muslims might actually think that limiting Muslim refugees into the U.S. is a good idea. That's not far-fetched.

And all Trump needed to do to win was a little better with minorities than Romney.

Saturday, November 19, 2016

House of Representatives Vote Update, Republicans 51.1%-49.9%

I've been busy compiling the House of Representatives vote also. This is more difficult, as some states don't update their votes. Other states don't list vote totals for candidates who run unopposed. Thus, I went to county election websites to get some of these. My numbers are incomplete, as every county doesn't list vote totals. When all the votes are counted, however, all the states except Oklahoma and Florida list vote totals for candidates who run unopposed.

Republicans now lead by 51.15%-48.85%, a drop from leading 52.1%-47.9% the day after the election. There are two main reasons for this. Because most of the unopposed candidates whose totals I've updated are Democrats they've gained quite a few votes from that research. California is the bulk of the post-election day vote counting and Democrats have gotten 64% of the vote in the state. Thus, I expect the margin to drop to roughly a 50.9%-49.1% Republican win. In 2012 Democrats won by 1.2%. So 2016 could be a 3.0% improvement for Republicans. I expect Donald Trump's margin improvement to be around 2.1%.

Presidential Vote Update, Clinton 47.8%-46.6%

The Presidential vote, linked here, is at nearly 133 million votes, 4 million more than 2012 and nearly 2 million more than 2008. The final number should top 136 million. Currently Hillary Clinton leads 47.8%-46.6%. Much of her gain has come from California. They've counted over 3 million ballots since election day and that's increased Clinton's national lead by nearly 1 million votes. I'm revising my final estimate to a 1.7% Clinton win. So Nate Cohn's 1.5% looks good but his 2.0% is a bit high.

California sticks out as an anomaly. Without California Barack Obama won the rest of the country 50.0%-48.4%. Without California Donald Trump won the rest of the country 47.9%-46.4%. That's a 3.1% swing in margin. The median was a 2.5% swing toward Trump and many of the margins were clustered there. While the rest of the country swung 3.1% toward Trump, California swung 3.5% toward Clinton. That strong California swing toward Clinton, along with ones in Texas, Massachusetts, and Arizona, has brought up Clinton's margin. These states will see Partisan Vote Indexes (PVI) change by 4-5 points from 2012 to 2016. In 2012 this only happened with small states. They tend to be more volatile.

Does this signal a dramatic shift in these states or just a Clinton-Trump anomaly? I think the latter. In Arizona and Texas Republicans did slightly better in the two party House vote than they did in 2012. In California Democrats are doing better than 2012 in the House vote but this can be the result of so many Top Two races without a Republican. I'll compare California district by district later.

Friday, November 18, 2016

I Won

The last three mid-term elections have gone poorly for the party in the White House. The two prior to those didn't, but they both had extraordinary one time events that don't get repeated. In 1998 Republicans impeached Bill Clinton and in 2002 George W. Bush was riding a post-9/11 popularity bump. The election prior to that, 1994, was a big wave for the out party. So is a Democratic wave inevitable for 2018?

Some people might conclude that Republicans are insulated because the mid-term electorate favors them. That's not exactly true. The mid-term electorate is more favorable to the GOP than the Presidential year electorate. That doesn't mean the electorate will be favorable to the Republicans. In mid-terms swing voters heavily swing to the out of power party, while the out of party power's supporters are energized to vote. A more favorable electorate didn't help the Republicans in 2006.

Right now Donald Trump isn't any more unpopular than he was when he was elected. Is his unpopularity assured? Based on recent history it appears that way, but there were reasons the parties were unpopular. In 2006 America didn't like Republicans due to Katrina, the Iraq War, and Republican corruption.

In 2008, Democrats got arrogant with their majorities. At first it appeared the two parties would work together on the stimulus. They were consulting with Republicans. In a meeting with the President Republican whip Eric Cantor objected to items in the President's package. President Obama told Cantor, "I won." He implied that since he won, America wanted what he was putting in the package, and that Republicans either went along with what he put in the stimulus or they'd be left out entirely. Republicans chose to be left out entirely and the stimulus passed without a single Republican vote.

During the next two years Democrats passed broad sweeping legislation that contained things Republicans wouldn't vote for. These bills included the Affordable Care Act, Obamacare. The Democrats thought that by passing bills Republicans wouldn't vote for would increase their popularity and make the Republicans look obstructionist. The opposite happened. Passing partisan bills galvanized Republicans against the Democrats and swung swing votes to the GOP. This energy was repeated in Barack Obama's second mid-term in 2014.

The lesson here is that Republicans and Donald Trump winning the election doesn't make them popular or mean America endorsed everything they wanted to do to the exclusion of the Democrats. If the Republicans arrogantly pass bills too conservative for Democrats to vote for them then they too will have a poor mid-term.

I'm not saying you can't pass any partisan bills. Barack Obama admitted later it would've been smarter to break up the stimulus into several bills, with one or more containing only tax cuts. Republicans would've been on board with a tax cut bill and the GOP voting for some of the President's bills might've softened opposition to him and might not have energized opposition that became the Tea Party. The lesson is to be cautious and judicious with the partisan bills that you'll pass along party lines. If not, Democrats will take the House in 2018.

Sunday, November 13, 2016

The Final Margin will be...

This morning Nate Cohn speculated that Clinton would win by at least 1.5%, if not 2.0%.

I guessed that even 1.5% would be stretching it. I was challenged by someone on Twitter that Clinton would get there because there were 7 million votes left in dark blue states. I went through the states and estimated there are 5.8 million votes left. They've already counted 4.0 million and I'd guess this person didn't know that. I also estimated 4.3 million of the 5.8 million are in dark blue states.

So I had to guess how many ballots are left in each state and how they'll vote. In many states the first part of that is based on a little info and I could be wildly off on states that haven't shared that data. I mostly used this estimate. In some cases, however, the states have surpassed the numbers listed already.

I estimated how the new ballots would go based on previous voting. That's pretty much a wild guess. I arrived at Clinton winning the remaining ballots by 20% and a final margin of 1.4%. For it to get to 2.0%, she'd have to win 7 million ballots by 29%. She only has margins that big in DC and Hawaii, so I doubt she'll average that for the rest of the ballots. So I'll predict the margin will be no greater than 1.5% and will likely be less.

Saturday, November 12, 2016

Saturday Morning Thoughts

This was the first time since 1986 that both parties entered the election with the exact same senate seats they won in the previous election. In both years there were no mid senate term specials. There was one appointee running in 1986 and two this year. Even more remarkable is that in both cycles no House seats had switched parties either in a special election. So both parties again had the same seats they held after the previous election.

I said that Trump would have a big improvement over Romney in Rhode Island and he did. I said he'd improve in Northeastern and Midwestern states and he did. I said that Clinton would have a big improvement over Obama in South Carolina and southern states. She didn't.

I said there'd be 140 million votes and there won't be. We are at 128.5 million and will surpass 2012's 128.9 million on Monday and 2008's 131.5 million by next weekend. The final number will be north of 133 million. I said Trump could win the electoral college while losing the popular vote and he did. I said the polls would be off and they were.

I said Democrats would win 3 senate seats. They won 2. They lost 2 I thought they'd win and 1 I thought they'd lose. I predicted Democrats would win 11 Republican House seats and Republicans would win 3 Democratic seats. Others predicted bigger Democratic gains. Democrats won 9 Republican House seats and Republicans won 3 Democratic seats. Eight of the 9 House seats Democrats won were on my list of flips. Two of the 3 seats Republicans won were on my list of flips. So I correctly predicted 429 of the 435 House seats.

Hillary Clinton won the popular vote but lost the electoral college. It’s worth noting that she won the popular vote because she improved on Barack Obama’s percentage of the two party vote in California, Texas, and Massachusetts. Clinton spent no money, didn’t campaign in, or had a GOTV effort in any of these states. She did worse in the states she did try in. Should she get credit for winning more votes when she didn’t have anything to do with it?

I've heard people say that Trump only won because Democrats stayed home. They point to Trump and Clinton both getting less votes than Obama and Romney. This is wrong in several ways. First, people were only taking the number of votes on election night and post-election night counting will result in an increase in total voters. Trump will pass Romney's total. Clinton won't match Obama's total, but there will be over 7 million 3rd party votes in 2016 than 2012. These weren't all Obama voters. Some Obama voters did die, leave the country, became infirm, or chose not to vote. Some Romney voters did too. Considering that Romney won voters 65+ 56% to 44% it's likely more of his voters died.

You have to look at where the votes were to see that Obama voters voted Trump. Dailykos, a liberal website, did an analysis and found Trump won 1/3 of the counties Obama won twice! If Hillary Clinton had held Obama's voters in Pennsylvania, Michigan, and Wisconsin she would've won the election.

Friday, November 11, 2016

Presidential Vote Spreadsheet

Get out your bookmarks. I've created a spreadsheet with the Presidential vote. This has been updated by checking the vote totals from the state elections divisions’ pages. Unfortunately, there are a bunch of states that don’t put their vote totals on the elections’ page. I assume they wait until the election is certified. So some of these states still have the election night numbers. Thus, I can't promise that this page will be 100% up to the moment everywhere but that it will be updated when I can.

We passed 127 million votes today. There have been nearly 4 million ballots counted since Tuesday night. We’ll pass 2012 some time early next week and the final tally will be larger than 2008. Right now Hillary Clinton leads by more than 400k votes, 0.3%. I anticipate that lead growing bigger, as California and Washington have a lot of outstanding ballots.

I'll update the House of Representatives page as more ballots are counted.

Thursday, November 10, 2016

Where Trump Helped Downballot

I said for months that Trump wouldn't impact downballot positively or negatively. Republicans weren't that associated with him that he wouldn't have positive or negative coattails. It seems I was wrong in rural areas. Trump may have gotten rural voters in Michigan, Wisconsin, Minnesota, and Pennsylvania who'd voted Democratic in the past to not only vote for him but also down ballot Republicans. He may have helped Jason Lewis win MN-2 and almost got the GOP MN-1 and MN-8. The congressional margins in rural Wisconsin races were laughers for the GOP. With himself as the VP candidate Paul Ryan won his district by 11. With Trump at the top of the ticket Ryan won by 35 points! I guess not fully supporting Trump didn't hurt him.

Did Trump drag Roy Blunt across the finish line in Missouri? Did he make the Ron Johnson and Todd Young margins as big as they were? I'm curious to see this election district by district, comparing 2016 to 2012. That analysis will be a few months off when all votes are counted and we can calculate the district by district performances.

CA-7 Ballot Update

While I haven't taken a deep dive into the California ballots yet, I do want to provide an update. I predicted Republicans wouldn't come close to taking any California Democratic congressional seats. I was surprised that Scott Jones only trailed Democrat Ami Bera by 1.2% at the end of election night. Most districts showed similar VBM and election day results. Here Ami Bera won VBMs 54%-46%. That was what I expected from the race. Jones then won election day ballots 53%-47%. A shift like that was uncommon and, to me at least, surprising.

There are 251k Sacramento VBM and provisional ballots left to process. This district should have more than half of them. For some reason Sacramento updated their totals with just a handful this morning. Scott Jones took these 222-161. The 61 ballots cuts Bera's lead 2,033. While that's not much, Jones getting 58% of the ballots would be huge if it were an indicator of how unprocessed ballots will go. He probably needs to get only 51% of what's left to win. What we don't know is how representative these ballots were. They may have been a Jones favorable precinct someone forgot to enter on election day. The post-election day count here should be watched.

Thursday Morning Thoughts

As of now three of Hillary Clinton's best states compared to President Obama are Massachusetts, California and Texas. Clinton didn't spend any money in any of these or have a GOTV operation in them. Clinton will win the popular vote but it wasn't because of anything she did to help win it. Donald Trump did better than Mitt Romney in the two party in almost every swing state. Hillary Clinton outspent Donald Trump in these states and, supposedly, had a better GOTV operation in them. The places she tried to win were the ones who narrowly lost.

In 2016 Democrats ran against Donald Trump the person. In 2020 they should/will run against President Donald Trump. It'll be a referendum on Trump as President, not as much about Trump as a person.

I keep hearing that turnout was low and that since Donald Trump got less votes than Mitt Romney he won because Democrats stayed home. Turnout wasn't low. As of now, turnout is down 3.5 million from 2012, but California is down 4.1 million. Washington state is down 933k compared to 2012. These two states have always counted a large number of ballots after election day. If you exclude California and Washington turnout is up 1.5 million. Over the next couple of weeks these two states will count all their ballots and we'll surpass the 2012 count. As of now, Texas and Florida are up 11%, Oklahoma up 9%, and Nevada up 8%. These states will also add ballots. Yesterday an additional 1.4 million ballots were added to the total and some states haven't begun counting their late and provisional ballots.

Wednesday, November 9, 2016

Republicans Winning House of Representatives Vote

I spent part of today accumulating the House of Representatives vote. You can find it here.

I’ll give a few caveats. One is that this isn’t the final vote. There are a lot of ballots to be counted, especially in California. That should help Democrats. The second is that I don’t have vote totals for 15 districts where a candidate ran unopposed. These votes are included in the final official canvass for every state but Florida and Oklahoma. There are 9 Democrats and 6 Republicans who ran unopposed and the vote totals haven’t been published. I expect we’ll eventually have them for 8 of the Democrats and 5 of the Republicans. That should help Democrats too.

The results are surprising. The meme has been that the Democrats had a big advantage in Presidential years and it was only going to get bigger. Yet the two party vote is 52.1%R-47.9%D. The 2014 vote was 52.9%R-47.1%D. That margin should narrow a bit when all votes are counted. This should dispel the idea that the Democrats have a majority coalition and Republicans will lose every election.

If Republicans did so well, why did the party only pick up three seats and lose nine others? Four of the nine Republican losses were in seats that were redistricted. As were two of three Republican gains. The other losses were mostly in swingy/light blue districts. Republicans already have most of the districts they'll retain with 52% of the vote. The party lost a few seats narrowly that would be unsurprising for them to pick up again in a future cycle. Considering that a couple of weeks ago people were saying the House was in play this isn't a bad result for the GOP. In two years they'll have a more favorable electorate. Mid-terms do usually go badly for the incumbent party and that might cancel out an electorate advantage.

First Wednesday Morning Thoughts

I've been saying for months Donald Trump could win because this is a deeply divided country ant that the Democrats were running a candidate as unpopular as Hillary Clinton. Most people dismissed that idea and predicted a big win for Clinton. I didn't think Trump would win but never dismissed the possibility he could win.

I'm most disappointed that some of my friends on Facebook have expressed such in intense level of hatred. That scares me and maybe says something about me. America will survive. We've survived every President up until now.

Can we finally do away with the idea that Democrats will win every election from now on and that "demographics are destiny?" We should probably wait until Democrats start winning Presidential and mid-term elections before we decide that's happening. I doubt Democrats will take the lesson they should from this election, "stop telling white people they're deplorable." They keep losing votes by alienating white voters.

Republicans appear to have done better in California than I thought but still could lose a few in the after election day counting. Yet their performance, compared to Republicans elsewhere, is disappointing.

Tuesday, November 8, 2016

California Predictions: Dems win big in Assembly, No change in Congress

I've written a lot of analysis about the races, so I'm just going to let my predictions, on the linked spreadsheet, speak for themselves. I predict no change in the congressional seats, but Democrats picking up five assembly and one senate seat. That'll give them supermajorities in both houses.

In 2012, I overestimated Republicans, but there were two reasons for that. It was the first year of top two, so I didn't have data to go on and there was a late surge of Democratic registration that no one knew about. Republicans didn't take all the seats I had them getting in 2014 but I had them in competitive races in a number of districts that people said were uncompetitive. In a few hours analysis and predictions won't matter. We'll see by the results.

Capital Weekly Exit Poll

Capital Weekly, a California politics publication, did an exit poll via the Internet of VBM voters. They didn’t weight the returns. That means that if there's a congressional district where the VBM returns were 35%D/45%R, their poll could be 50%D/25%R.

They don’t give a Presidential bottom line but it looks like Hillary Clinton 66%-24%. Barack Obama won 61%-38%, so Clinton would nearly double his margin. California is roughly 10% of all votes, so it heavily influences the national vote. If Hillary Clinton were to win states by 19% more than Barack Obama she'd win the election by 23%. There isn't a lot of room for Clinton to improve on the California minority vote because Barack Obama did so well with it. She doesn't do as well as he did with African-Americans and Asians, but does better with Latinos. Overall she's only helped a little bit. Where she picks up the vote is the white vote. Clinton wins that 63%-27%. Who said Hillary Clinton was weak with working class whites? Mitt Romney won it by 53%-45%. That's a 44 point swing. If we apply that to the whole country, which has a much whiter electorate than California, Hillary Clinton wins the national popular vote by 32%. I'm guessing those saying it's a 4 point race will have egg on their face!

Clinton would heavily Republican districts like CA-4, C-22, CA-39, CA-42, CA-45, CA-48, CA-49, and CA-50. She’d win Republican congressman Ed Royce’s district by 23 and congresswoman Mimi Walters by 27. Trump would win only 4 congressional districts, CA-1, 8, 21, and 23. Strangely, one of those districts is David Valadao’s district. Barack Obama won that district and Valadao is considered to be in trouble.

Needless to say, the Republican congressional delegation would get wiped out. Jeff Denham, Darrell Issa, and Steve Knight are considered to be in competitive races. Not only would they lose but so would Dana Rohrabacher, Ed Royce, Mimi Walters, and Ken Calvert. Duncan Hunter wins a squeaker. I’m guessing none of them will see it coming.

Going further down ballot the GOP would lose 3 state senate districts and 12 assembly districts. If you're a Democrat reading this, and those sound about right, don't bother reading my California predictions coming in the next post.

Monday, November 7, 2016

California and the Latino Voter Surge

There's a lot of talk about a Latino voting surge this year and a big question is whether it exists in California. Looking at the PDI information we see that in 2016 Latinos have made up 15% of California VBMs, up from 12% in 2012. That'd appear to be a surge but they were 17% of the voters who got a VBM ballot in 2012 and are 22% this year. So the percentage of Latinos returning their VBM ballots has actually dropped from 70% to 68%. Because more people are receiving VBMs there's a smaller pool of election day voters than there were in 2012. We could see more election day VBM returns than in the past, but if it's the same then we might not have much, if any, increase in Latinos as part of the California electorate.

That's not to say that there isn't a Democratic surge. There is a strong one, but I think it's largely white Democrats who are driving the increased VBM return, with Latinos secondary. It's also worth noting that the increase in Democrats isn't as good as it seems. In 2012 the VBM electorate was 43%D/32%R and the returns were 43%D/36%R. This year the VBM electorate is 45%D/27%R and the returns are 46%D/32%R. So Republicans are still returning at a greater rate than Democrats. That's the end of the good news for the GOP. In 2012 the overall electorate was 44%/29%R and now it's 45%D/26%R. The Democratic surge in VBM ballots might not mean more of a Republican election day vote compared to the past.

2016 Presidential Polls 11/7/16

Here is the last national v. state poll comparison. While Hillary Clinton ticked up in the national polls over the weekend from a 2.7% lead to a 3.3% lead, she didn't see a corresponding jump in the state polls. They remained at 3.1%. Some of this could be due to the presidential polls catching up to the state polls, but more likely it was because a lack of state polls, especially in some bigger states, didn't move the needle.

We did see a tightening in swing state polls. Going into the weekend Clinton led these by an aggregate of 1.0% and now leads by 0.4%. That suggests a closer race with around a Clinton lead of 2.7%. If the results end up like the spreadsheet Hillary Clinton wins 316 electoral votes to 222 for Donald Trump. Trump is polling within 2% on another 48 electoral votes, however. That'd give him 270, enough to win. There are another 34 electoral votes where Clinton leads by roughly 3%. With a little tightening and a polling error in Donald Trump's favor and he could win the presidency even when losing the popular vote by 2%. The polls underestimated Democrats in 2012 and Republicans in 2014 and I don't have great confidence they're getting the electorate right. Of course if the polling error is in Clinton's favor she could top 360 electoral votes.

Sunday, November 6, 2016

National Predictions

A lot of people are going to give you a prediction on the Presidential election, so you don't need one more. I do want to note that 538 has Trump ahead in states that total 265 electoral votes despite being behind by 2.9%. He needs 270 to win. They show Trump with 245.7 electoral votes because they have Trump winning the three closest states by 0.2% or less. That seems unlikely. Trump would have to be extremely lucky to win if he loses the popular vote by more than 2.0%. He may have an advantage in the swing states so that he doesn't need to win the popular vote to win the electoral college, but not while losing the popular vote by that much.

In 2012 there were 129,237,642, down from 131,473,705 in 2008. That was a surprise because there were more American 18 years or older in 2012 than in 2008. I predict that we'll see a good bump in 2016 with the electorate over 140 million. Turnout looks strong in early voting and there appears to be enthusiasm on each side.

Donald Trump will improve over Mitt Romney in Northeastern and Midwestern states where Barack Obama did well with working class white voters. Some of these voters will turn to Donald Trump. I predict that Trump will see his biggest in Rhode Island. Barack Obama won the state by 27.5%. Donald Trump will lose the state by double digits but the gain will be big. Other states where Trump will make big gains are New York, Maine, Iowa, New Jersey, Connecticut, and Michigan. Unfortunately, the only state Trump might win is Iowa, although he has a shot at Michigan.

For years Republicans have been winning 80-90% of white voters in southern and some western states. There are no working class white voters voting Democratic. But there are plenty of suburban white voters voting Republican. Trump is a turn off to some of them. In other states working class whites might offset are even more than offset losses amongst these voters. South Carolina, Arkansas, Idaho, Georgia, and Tennessee should be states where Hillary Clinton will make the biggest gains. She is unlikely to win any of these states but has a chance at Georgia.

Don't be surprised if polling is off. While polling was pretty good in the primaries it wasn't in the 2012 and 2014 elections. The electorate this year is likely to be different than previous ones. Will there be a Latino surge? If so, how much of one? Are the missing white voters working class Trump supporters who'll show up this year? I'm not going to predict whether the polls are wrong for Republicans or Democrats but that they are off.

The state expected to be the tipping point state won't be. So many states are polling within 4% that no one can predict which one will be the closest. There are some early states to watch, however. Ohio, Florida, Georgia, Virginia, and South Carolina all close at 7 or 7:30 PM eastern. (Part of Florida closes at 8, but the networks might not wait there) If the margins are better for one candidate in these states then that candidate might be surprising. Exit polls are usually, but not always accurate, and don't get fooled by vote counts like Karl Rove did in 2012. Mitt Romney was close enough to win Ohio when Fox called it for Barack Obama, but a disproportionate number of the remaining votes were from the city of Cleveland. Obama ran up big margins there and he did win comfortably.

The Senate isn't always a sign of a wave, but the House is. Every cycle there are expected to be around 10 states that are close. If one party wins a good percentage of them it may be because of good candidates or other state circumstances. The sample size is so small that a few breaks could turn a 5-5 split into a 9-1 advantage. I'm predicting that Wisconsin, Illinois, and Pennsylvania flip to the Democrats and the Senate will be 51-49 Republican. I wouldn't be surprised, however, if either party does several seats better.

The House is 435 elections. Some of them are going to be big romps no matter what but environment impacts most of them too. A candidate that'd normally win 67%-33% might win 64%-33%. So while the Senate might go big one way or the other the House definitely won't. I'm predicting the Democrats pick up 11 seats, FL-10, FL-13, IL-10, MI-1, MN-2, NH-1, NJ-5, NV-3, NV-4, TX-23, and VA-4. Republicans will pick up 3 seats, FL-2, FL-18, and MN-8. That'd net Democrats 8 seats and put the Republican advantage at 239-196. Democrats might do better but even if they do, they won't exceed 205 seats.

Saturday, November 5, 2016

2016 Presidential Polls 11/5/16

Polling yesterday kept my state by state tracker pretty much the same. Hillary Clinton went from leading by 3.2% to 3.1%. On the other hand, the 538 national margin dropped from 3.1% to 2.7%. One of the reasons the state polls didn't drop by much was that a new California poll came out showing Clinton leading by 24 points in the state. That upped her margin there to 22 points, a 2 point lift. Because of its size, California accounted for 0.2% for Clinton. Without it, she would've dropped to 2.9%. On the other hand, the swing state moved a little bit toward Clinton and those now suggest a 3.3% Clinton win. If Clinton can keep the polls in the 3 point range, she probably wins even with a poll bias toward her.

Friday, November 4, 2016

2016 Presidential Polls 11/4/16

While I'm still getting the percentages on many states from Internet only polls my state aggregate has come in line with the national averages. The state polls show Hillary Clinton with a 3.2% lead. The national polls show her with a 3.1% lead. Clinton led by 3.2% on Wednesday, so she's stemmed the bleeding in the last few days. If Clinton can keep her lead above 3.0% in polling she should win even if there's a polling error.

My current projection has Clinton winning the electoral college 319-219, but Donald Trump is polling within 4.0% in states with 109 electoral votes. Trump would only need 51 electoral votes to win. Contrary to what you may be hearing, if Trump gets the race closer he has a number of paths to victory. Michigan, Pennsylvania, North Carolina, Colorado, and Virginia are all in play.

Senate polling isn't always moving in the same direction as Presidential polling. Trump seems to be gaining Never Trumpers, people who were likely already voting for Republican Senate candidates.

Thursday, November 3, 2016

California Field Poll

When polling California correctly you use a voter file and poll 4s and 5s. These are people who've voted in either 4 out of the last 5 elections or all 5. Expected turn out should determine how many lower propensity voters you include. If you know what you're doing you can predict turn out by party registration, as well as by ethnic group, age, and gender. If you don't understand that you end up with a 32% Hispanic/29% 18-34 electorate the way SurveyUSA did or a poll with a D+30 electorate the way Sacramento State did. And you're less likely to get Paul Mitchell thinking about the after life.

The Field poll is the gold standard. They nailed the Brown-Kashkari result in 2014. Right now Donald Trump is polling at a slightly closer margin nationally than Mitt Romney did. Because California is roughly 10% of all votes, it closely correlates with the national changes. If Hillary Clinton were to win by the 36 points Sacramento State had in their poll, she isn't going to win nationally by a smaller margin than Barack Obama. He won by 23 points here.

Field has Clinton with a 20 point lead. They put the already voted margin at 23% and have Trump doing better with later VBM and election day ballots. This is a reverse of past California voting patterns, as Republicans tended to do better in early voting. It’s more in line with Democrats voting heavier early nationwide.

If Field is right, then Trump is likely to do better in individual districts than Romney did in 2012. Trump is likely to win districts Romney won and maybe one or two he didn't. That runs contrary to the media meme that Trump would get slaughtered in California, especially in swing districts. It's likely that Trump will improve on Romney a bit more in Whiter districts and may lose some ground in heavier minority districts but if Clinton wins the state by a similar margin as Barack Obama did, voting patterns should be similar to 2012. Democrats have been trying to tie Donald Trump to swing district Republicans. If Trump has similar results to Romney, that won't work.

Wednesday, November 2, 2016

2016 Presidential Polls 11/2/16

In the last day Hillary Clinton's national lead has dropped by roughly 0.5% from 3.7% to 3.2% and her lead in the state polls has dropped from 4.7% to 4.3%. In the 14 swing state polls Clinton is 1.5% behind Barack Obama. If that margin were applied everywhere Clinton would be up by 2.3%. As I've mentioned before, the non-swing state poll data I'm using is older and reflects a time when Clinton had more of a lead. Based solely on the polls I'm inclined to believe her lead is 2.5%-3.0%. That's very close, especially when there appear to be more swing states than there have been in the past. Donald Trump has a number of paths to 270.

As I mentioned yesterday in my senate polling analysis, polling has been inaccurate the last two cycles. Barack Obama won by 3% more than the polls said he should. If the polls are wrong, Trump could be in the lead or Clinton could have a bigger lead. The polls are close enough where polling error, if it happens, could come into play.

Tuesday, November 1, 2016

CA-49: Trump tells supporters to vote for Darrell Issa

Donald Trump did something today that down ballot Republicans need and he hasn't done, tell his supporters to vote for them.

Some of Trump's supporters don't like Republican politicians and would be just as happy if they lost and there was a Democrat in the seat. They want to punish Republican politicians for disloyalty. If Trump tells them to vote for someone, they're more likely to vote for that candidate. If Trump tells his supporters to vote for down ballot Republicans across the board Republicans should do very well. The Republican voters who supported Issa in the past who might vote Clinton this time should still support Issa even if Trump is endorsing him. They are still Republicans. Aaron Blake at the Washington Post thinks this is bad for Issa.

The media, at least sometimes, is accepting that Hillary Clinton is winning the race by a similar margin than Barack Obama won in 2012. Right now it looks like less. Yet they act like Trump is going to get killed in swing districts. Mitt Romney won this district by 7 points. If Trump loses by a similar margin as Romney did, his votes have to come from somewhere. They're likely to come in similar margins as Obama-Romney had. Yes, Donald Trump is very unpopular in California and will get killed here. But Matt Romney got killed here and still won CA-49.

Mitt Romney won the district with 153,856 votes. Issa won by nearly 10% more but only got 159,725 votes. That's because roughly 20,000 people who voted for President didn't vote in the congressional race. Issa would've won with the number Mitt Romney had and would win this election if he gets the same number of votes Donald Trump gets here. I think this is a big positive for him.

2016 Presidential Polls 11/1/16

According to the state polls, reflected on my polling spreadsheet, Hillary Clinton's lead has dropped from 6.1% to 4.7% since Friday. I think the race is closer than that. The poll analysis by 538 puts Clinton's lead at 3.7%.

On Friday my state polls were 0.6% above national and now 1.0%. Why do I think the national is closer? There are 14 heavily polled states. 538 lists 16 states on the side of their national page, but I excluded Maine, Minnesota, and Utah because they're not heavily polled and included Missouri, which has been. For my election result in most of the remaining 37 states and DC I'm using Internet only polls with an occasional traditional poll in there. In some cases my traditional poll is older than swing state polls and the I really question the accuracy of Internet polls.

So I totaled the vote if only the 14 most commonly polled swing states were used. Clinton leads in them by 1.1%, compared to Barack Obama winning them by 1.7%. Obama won by 3.8% in 2012. If Clinton is doing 0.6% worse she has a 3.2% lead, a bit worse than the 3.7% lead 538 gives her. I don't think it's that low, as Trump seems to be doing better in swing states than in non-swing states. It's conceivable that Clinton could win by less than Barack Obama did in 2012 but win by more in California. If Clinton wins California by 3.0% more her overall margin moves roughly 0.3% toward her. If she wins the election by 0.6% less than Barack Obama, that means Trump needs to make up the 0.9% in other states.

A number of these swing states appear tight. If Trump chisels Clinton's lead by even 2% more, he could conceivably win while losing the popular vote.