Silver on Clinton’s Firewall
Steven L. Taylor
Friday, November 4, 2016
To add to the previous post, Nate Silver:
The point, as we’ve said before, is just that Clinton’s so-called firewall is not very robust. If you’re only ahead in exactly enough states to win the Electoral College, and you’d lose if any one of themgets away, that’s less of a firewall and more of a rusting, chain-link fence.
About Steven L. Taylor
Steven L. Taylor is a Professor of Political Science and a College of Arts and Sciences Dean. His main areas of expertise include parties, elections, and the institutional design of democracies. His most recent book is the co-authored A Different Democracy: American Government in a 31-Country Perspective.
He earned his Ph.D. from the University of Texas and his BA from the University of California, Irvine. He has been blogging since 2003 (originally at the now defunct Poliblog).
Follow Steven on Twitter
She’s comfortably ahead in enough states to guarantee a win. She has small leads in enough states to win in a blowout.
I’ve been saying this for a while. The idea that the electoral college has some “firewall” that will magically protect Clinton is nonsense. We went through this a month ago. Clinton’s national poll numbers fell. Everyone said, “but her firewall!”. And then the firewall collapsed. And suddenly the odds were dead even. As Silver points out, many of the firewall states have been sporadically polled. So there could be some nasty surprises come election night (remember Michigan in the primaries).
You CAN win a national election without winning the popular vote but it’s only happened twice in our history. And even if somehow Clinton did win, what kind of presidency would she have? No, if she wants to win, she needs to win. Firewalls will not save her.
A small ray of light in the dark:
A Vote for Good
538 currently gives Hillary a 66.5% chance to win. That’s far from a “guarantee”, You have a 1/3 chance of waking up on Wednesday to “President-elect Trump”.
Two observations: 1) it is more likely that Trump would lose the pop vote and win the EC than the other way around, and 2) the last president who lost the popular vote and won the EC by a hair went on to be a two-termer.
@Kari Q: She is in the lead and should be favored to win. But nothing is guaranteed.
@C. Clavin: Sadly, not enough to be confident of a specific outcome.
I agree that from a dispassionate point of view one has to conclude that Clinton is the more likely winner. The problem is, a) the stakes here make dispassion difficult (although not impossible), and b) even from a dispassionate point of view, the trends lines and margin of error make the outcomes far from certain.
Well, I’d rather wake up to a Hillary Clinton presidency than a Donald Trump presidency, even though I know full well that Republican leadership has recently indicated that they will not consider Hillary Clinton to be a legitimate president, and are seriously considering opening more investigations as a prelude to impeachment proceedings.
Republicans have not considered the previously elected (each to 2 terms) Democratic presidents to be legitimate – they impeached Bill Clinton, and indulged a racist Birther movement for most of the Obama administration. This is what Republicans do.
@Steven L. Taylor: This is true, which is why GOTV is so important.
Also why I’ve actually been slightly relieved to see 538 showing a tighter race – when her chances were at 90%, I was worried that a lot of people would think “it’s in the bag” not bother with long lines and inconvenience of getting to the polls (esp here in PA were there’s no early or absentee voting unless you’ll be out of town, in a nursing home, etc.)
With a real chance that we get stuck with that Putin-loving, xenophobic, sleazy con man, I think a lot more people will be motivated to get and vote.
This is another example of how everybody points and laughs at how screwed up Republicans are while dysfunction on the Democrat side is right there for everyone to see. This firewall business fits in with all the demographic delusions that have really flowered since 2012.
Democrats have been a minority in the House of Representatives for 22 straight years. Gerrymandering, you say? Then why have they spent almost all of that 22 years as a minority in the US Senate and, I believe, most of it controlling a minority of state legislatures and a minority of governorships? I remember seeing a segment on MSNBC a few years ago on the supposedly bleak electoral future of the GOP and it was a New York Times person who pointed out that if the Democrats lost the White House, it would be their future as a national party that would come into question.
@Steven L. Taylor:
Tell me about it.
And the fact that the FBI is in the tank for Trump has me more concerned than I was before.
Imagine a thin-skinned vindictive man-child like him with the FBI at his beck and call. I mean…it sounds unbelievable. But a week ago the idea of the FBI trying to influence the election was unbelievable too…
“You CAN win a national election without winning the popular vote but it’s only happened twice in our history”
Four times (1824, 1876, 1888 and 2000), but who’s counting?
I am becoming very confused by Silver’s methodology. It seems any single poll which is more than likely an outlier tends to affect the odds more than it should. This morning a Georgia poll had Trump only winning by 2%, and Hillary’s odds went up over 1%.
If you are relying on GOTV to beat Donald Trump, you’ve already lost.
Maybe…just maybe…you should have been a little concerned about the Democrats nominating someone to run for President WHILE they were under FBI investigation for possible criminal violations.
And just to clear something up, is it now okay for right wingers to question the fairness and integrity of Comey for his refusal to recommend an indictment against Hillary? Or are only liberals allowed to impugn his character when he does something they don’t like?
The Democrats had a majority in the 110th (2007-2009) and 111th (2009-2011).
We are going to find out both a) what kind of efforts the two campaigns have constructed, and b) how much it actually matters. The standing view is that it matters.
And contrary to MBunge’s assertion, campaign normally do rely heavily on GOTV efforts.
@MBunge: Comey has unnecessarily and inappropriately inserted himself into electoral politics, including his original public pronouncement back in the summer.
He has made himself part of the political calculus, which he should not have done.
Silver’s model tries to combine state polling and national polling to get a global picture. Because state-polling data can be fickle, you try to use national and other state polls as a counter-weight. While this is, ultimately, fifty-one related elections, they aren’t independent elections. A national fade or rise will show up in state voting, one way or another. That’s why this talk of a firewall was always nonsense.
If Trump is only leading Georgia by two points (1) that means that there is a better chance Clinton could win Georgia; (2) that may indicate that he’s weaker nationally than the current analysis says. It’s weighted appropriately by the state, the pollster, the sample, etc. But the general idea is that if Trump is only leading Georgia by two points in a reasonable-quality poll, that’s not a good sign for his campaign.
No, I don’t think it’s ‘nuf ced. That analysis assumes that there is minimal bias in the polls, which has frequently shown up in past elections (e.g., Obama outperformed the polls by a couple of points in 2012). I think the Princeton analysis WAY underestimates the chance that the polls are globally off by a couple of points. IF they are, this election will be much closer. I think it was pointed out in another thread that Silver gave Reid a 1/6 chance of beating Angle while Wang gave him none. This is the reason I prefer Silver: he’s way more conservative. Wang has an advantage in more stable and less prone to the up-and-downs of polling.
My head tells me that Clinton will win this by a few points and maybe 50 EV. There is as much a chance of a Clinton landslide as a Trump victory. But my gut tells me it’s not over until the actual votes are counted. And I’ve seen way too many sure things end in tears.
Have there been any studies on how early voting affects state-level polling quality? I am not in an early voting state (NH), but once someone has voted and is done with the process, does this impact the likelihood that they will respond to surveys?
I can reason out either way: either people would be more inclined to respond to say who they voted for and that it’s done, or, people, once done with voting, decide they are no longer answering the phone. I don’t know that it would impact results necessarily, but could affect the overall pool of respondents, thereby driving up the MOE.
Especially on the Democrat side, which make the Philadelphia transit strike all the more interesting. Basic union tactic to strike when the employer needs you most to increase the leverage. Now the unions are striking when Democrats who run the city most need the GOTV to satisfy their national DNC masters. Somebody’s going to get a big pension increase, except to pay it, the city will have to cut back on public transportation.
@MBunge: “Then why have they spent almost all of that 22 years as a minority in the US Senate ”
Do you really not understand that the Constitution set up the Senate to favor rural, less-populated states at the expense of the larger urban ones? That the 600,00 people in Wyoming or the 700,00 in North Dakota are represented by the same number of senators as the 38 million in California or the 20 million in New York?
I know that you are on some weird quest to be the Only Moral Arbiter on the internet, but if it requires you to pretend not to understand the basic facts of our governmental system, what’s the point?
That’s not fair. Trump doesn’t love Putin. It’s purely a business relationship. Like Melania.
@Moosebreath: Silver’s methodology goes a little wacky when there are a large number of undecideds, and a significant number of voters for third parties. I cannot say that is wrong.
@wr: I overlooked the Senate comment.
The Dems controlled the Senate as recently as the 113th (2013-2015). Indeed, the Dems controlled the Senate in the 110th, 111th, 112th, and 113th (i.e., a pretty hefty chunk of the 22 years MBunge noted).
In the first statement, Comey had no choice as his boss had put him on the hotseat with her exparte meeting. So he laid out the facts of the case and then recommended no prosecution as she needed. The no prosecution was for two reasons. As James has written, the rule of law does not apply to high government officials who can count on prosecutorial discretion (hopefully, only for malum prohibitum crimes and not murder and such). And, the email investigation was sandbagged from the beginning as the DOJ would not impanel a Grand Jury for the investigation so the fix was in. But this fix also gave Comey leave to lay out the facts as there would be no prosecution.
The letter to Congress was probably forced upon him by the circumstances of the Weiner case. There was already a Grand Jury called for that case so evidence would start being entered into the record. Plus, the investigations into Weiner’s wiener was not just federal as state law enforcement would be involved and as such state agents beyond the control of the DOJ cover up may have had knowledge of the email find. Not content, but the metadata and existence as part of the sexting a minor investigations. So the information was likely to get out and withholding the information would also be inserting himself into the election, and the likely revelation by other than the FBI would have been more disrupting to the election.
Dr K has a good piece in NYT this morning.
Yes, way better to have Hillary, even if she’s facing a GOP House and Senate than Donald Trump. My expectation is that President Trump would spend his days exacting revenge for petty slights, trying to monetize the office, and rubber stamping anything Ryan and McConnell sent him.
Keeping hope alive!
@Steven L. Taylor:
I agree. I also agree that if she continues to decline in the polls, she will lose the small leads she has in states like North Carolina. I will further say that the “blue wall” is more about, all other things being equal, the Democrats have an advantage in the electoral college vote. An advantage is not a guarantee, of course.
Silver’s analysis this year is simply seeming off to me. He ignored the polls in the Republican primary and got it wrong. Now his model is adjusting polls to every slight swing and hiccup in national polls, and producing a pretty volatile graph for an election that simply hasn’t been that volatile. I’m grumpy with Silver, because he ought to be better than this.
It really drives me crazy that the 625,000 people in Vermont get the same representation in the Senate as the 27 million in Texas.
MBunge, don’t waste your breath scolding the neolibs on here for their support of a corrupt grifter.
Its just like debating the W slurpers back in 2002 over the insanity of the Iraq war and WMDs. Their eyes glaze over, mouth foams, mind shuts down….then the defense of their support for the ridiculous begins.
Of course you made the correct point, “gee, lets nominate an ethically challenged, most unfavorable Democrat to ever run for POTUS, all the while hanging over head is an FBI investigation which could result in charges brought against her. My Gosh, what could possibly go wrong?”
Oh, how I love the down votes. Supports my whole contention of the extraordinary failure of the modern lib Dems to win elections.
You are right – the GOP has waxed the Dem party in all phases of electoral politics, save POTUS.
The GOP has a record number of state houses, Congresspersons and Governors.
And then we get WR to lecture on the nature of Senate seats somehow preternaturally going to the GOP because of the rural/founding fathers/bias as he pretends to understand the basic facts of our governmental system
Hey wr please explain to me you pedantic doosh, how those small rural states somehow were Democratic from 1932-1982 since we have the exact same electoral system in place?
Maybe, just maybe, the neolibs have zero clue on how to attract middle class and white working blue collar voters anymore as they tilt to the DLC moderate corporatist elitists.
In my day, college educated white voters always voted GOP and the hard hats, the blue collar votes always went Democrat. Now, its completely reversed.
Somehow we lost all of that and we can’t blame it on the “stupid” voter.
“If Trump is only leading Georgia by two points (1) that means that there is a better chance Clinton could win Georgia; (2) that may indicate that he’s weaker nationally than the current analysis says. It’s weighted appropriately by the state, the pollster, the sample, etc. But the general idea is that if Trump is only leading Georgia by two points in a reasonable-quality poll, that’s not a good sign for his campaign. ”
The word “if” in both places and the word “appropriately” are doing some heavy lifting there. Silver’s methodology seems to be too susceptible to a single outlying poll, treating it as showing the correct trend country-wide.
I don’t see why. It didn’t stop the GOP.
Trump is either under investigation or actually facing charges for fraud, racketeering, labor violations, various charges relating to the Trump foundation, including bribery, tax evasion, and illegal campaign contributions, and, oh yeah, alleged child rape.
Seriously, the issue HRC was being investigated on was mostly BS, as Matt Yglesia lays out here:
The real Clinton email scandal is that a bullshit story has dominated the campaign
@Gavrilo: I agree. Is that supposed to be some snarky little accusation of hypocrisy on my point? Because as usual, you fail.
Humph…certainly those circumstances didn’t hurt George W. Bush’s presidency…
Two points…first, Democratic losses in down ballot races have been the most severe under President Obama…a credible argument has been made that he bears some responsibility for that…even his biggest fans need to acknowledge this issue…and second, as long as the Republican Party is openly hostile to ethnic minorities, as well as most other minorities, and holds Luddite views on many scientific issues, the Democratic Party will be very competitive as a national party…
@Moosebreath: “Of course you made the correct point, “gee, lets nominate an ethically challenged, most unfavorable Democrat to ever run for POTUS, all the while hanging over head is an FBI investigation which could result in charges brought against her. My Gosh, what could possibly go wrong?”
Gosh, Q, you’re right. We Democrats should never nominate anyone who is not first approved by Republicans on House committees and the membership of the FBI. Because clearly they only want us to have a strong nominee.
@the Q: “Hey wr please explain to me you pedantic doosh, how those small rural states somehow were Democratic from 1932-1982 since we have the exact same electoral system in place?”
See, there was this little thing called the Voting Rights Act, and all of a sudden the Democratic Party was taking the side of “those” people instead of the upstanding patriots who only wanted to enforce Jim Crow laws.
I guess you’ve forgotten all of this by now. Doesn’t seem like you have anything left in your head except rage.
@the Q: “Maybe, just maybe, the neolibs have zero clue on how to attract middle class and white working blue collar voters anymore as they tilt to the DLC moderate corporatist elitists.”
And there’s the tell. “White working blue collar voters.” The only people who matter to the Qs of the world. This is the whole problem with Democrats to old creeps like him — the Ds stopped putting down minorities to raise up good ol’ real American white folk.
Your time is over, old man. We are a rich, proud, diverse nation. And if you want a party that will put down minorities to help members of your own race, it’s time to join the Republicans. Or the Klan.
WR. “We Democrats should never nominate anyone who is not first approved by Republicans on House committees and the membership of the FBI. Because clearly they only want us to have a strong nominee.”
What kind of horsedung is that? Sarcasm yes. Total inanity. Ditto.
The Dems had a clear choice to get a quasi socialist in office, and thanks to the wiki leaks, we see the DNC doing everything they could to thumb the scale in favor of Hillary. Surely, no one here will question that bias? The CGI corruption. The hundreds of miilions the grifters “earned” trading their public positions for filthy lucre etc. Bernie had none of that baggage to carry. And the mood of the country right now favors some “socialist” solutions. Huge chance lost because of the corrupt Clinton machine gaming the system in her favor.
And now, we see she may well lose to an infant lunatic.
Way to go, neolibs, way to go.
That must be why Republicans have done so well in the past few years…because so many people want “socialist” solutions…
The thing is yes, he ignored the polls in primary and that was a mistake.
Now he is fully in the polls and analyzing them in the same way as the last two cycles. As such, I am not sure that there are grounds to be grumpy save that the analysis is not producing the answers one wants.
And trust me: I prefer Wang’s position to Silver’s at the moment.
I totally agree. There really is no good justification for the disparity in question.
Representation in the Senate was born wholly out of a political compromise, not some deeply thought out political theory (also at a period of time in which there were not the kinds of population disparities that we currently have and have had for a long time).
My overarching concern is not defending Clinton, it is avoiding what will be true disaster in Trump.
I was just pointing out that you overlooked the vast population differences among the states often benefit Democrats in the Senate. In fact, in the 10 most populous states according to the 2010 census, 10 of the 20 Senators are Republicans. And, in the 10 least populated states, 10 of the 20 Senators are Democrats (that includes Angus King and Bernie Sanders.)
So, while I understand your desire to discount the GOP Senate majority as an indicator of the strength of the Republican Party, it’s a false assumption.
@Steven L. Taylor: My overarching concern is not defending Clinton, it is avoiding what will be true disaster in Trump.
I’m not so far gone to deny that a President Trump could be pretty darn awful. Most of the reasons for that are well known, though an overlooked one is that Trump would likely fill quite a few positions in his administration with cronies, flunkies and assorted hangers on that have little to no background in government service, resulting in one scandal after another.
I do think this hysteria over a Trump disaster is a text book example of class bigotry. For example, Trump throws around a lot of bellicose rhetoric but this is 2016, not 1016 or just 16 AD. We’re not getting into a war because President Trump has a Twitter feud with the Prime Minister of Belgium. Trump’s actual espoused policies, on the other hand, are more dovish than any major Presidential candidate since Eugene McCarthy in 1968. And if Trump is as easy to manipulate as all of his critics say, the real risk is foreign leaders will flatter Trump into doing exactly what they want.
Hillary Clinton, on the other hand, has employed her fair share of bellicose rhetoric in the past. She has a track record of supporting disastrous military and foreign policy mistakes. And she is on record, right now, supporting policies that are far more likely to get the US into another war than anything Trump has said.
@wr: We Democrats should never nominate anyone who is not first approved by Republicans on House committees and the membership of the FBI. Because clearly they only want us to have a strong nominee.
Hillary Clinton’s negatives were higher than her positives months before the Iowa Caucuses. The fact she was being investigated by the FBI for possible crimes was known months before the Iowa Caucuses. You ignored those facts and the suggestion that the Democrats couldn’t have found a candidate who didn’t suffer from both of those weaknesses is disproved by the other fact that the Democrats HAD another potential candidate who didn’t suffer from either of those weaknesses.
BTW – new polls today showing very close races in New Hampshire, Pennsylvania and Michigan. So … as happened before … the state polls are showing what the national polls are. You can’t depend on gaming the EV system. You have to win.
Because there are a lot more small rural states that go Republican than there are big states that go Democratic, you drooling utter moron. Because the 40 million people in California get the exact same number of Senators as the 600,000 people in Alaska. Because the combined populations of Idaho, Wyoming, Montana, Nebraska, North Dakota, South Dakota, and Iowa still don’t equal the population of California but produce six governors to California’s one.
Yes, I agree. It should drive you crazy.
For example, Trump throws around a lot of bellicose rhetoric but this is 2016, not 1016 or just 16 AD. We’re not getting into a war because President Trump has a Twitter feud with the Prime Minister of Belgium.
For example, Bush throws around a lot of bellicose rhetoric but this is 2000, not 1000 or just 20 AD. We’re not getting into a war because President Bush has a family feud with Saddam Hussein of Iraq.
That’s just not true. Of the 10 most populated states, 7 currently have Republican Governors and of the 10 least populated states, 5 have Democratic Governors.
@MBunge: “I’m not so far gone to deny that a President Trump could be pretty darn awful”
And yet you’ll vote for him.
Because he doesn’t have a vagina.
I first saw the headline as “Sliver In Clinton’s Firewall.”
And I think “Silver Sees Sliver in Clinton’s Firewall” would have been an even better one.
No, it’s key to Bayesian thinking. You never throw away data. You try to find the most probable model that reproduces the data you’re seeing. Silver is constantly onto pollsters to publish their outlier polls so that the error analysis can be improved.
@MBunge: While I will agree with you that Trump is more dovish than Hillary (not a high bar to jump, by the way), I find the story that during his intelligence (or lack thereof in his case, apparently) briefing he asked something to the effect of why can’t we just nuke them not one but several times outweighs any potential advantage that we might get from his non-intervention postures–and they may BE just posturing.
Let me try to explain my annoyance by pointing to another example.
This week, there was an unequivocally good poll for Clinton: the Marquette poll of Wisconsin showing her leading by 6 in Wisconsin. This was a small improvement for her (up from +4 in the previous), it came at a time when people were worried that her position in the Midwest was crumbling and showed it wasn’t.
Marquette is the premier pollster in Wisconsin. What Field is to California and Monmouth is to New Jersey, Marquette is to Wisconsin. This was great news.
At 538, that poll decreased Clinton’s odds of winning. It is also ‘adjusted’ to Clinton +4. There is no reason to make adjustments to Marquette. They do not have a history of bias or lean. They aren’t partisan. They are as good as a pollster gets. And they show her ahead by 6 less than a month from the election. And that’s bad for Clinton?
The change in state polling is expected. State polling follows national polls, and national polls tightened at the beginning of the week. It looks like the national polls have begun to rebound for Clinton, but there may not be time for state polls to pick up the change, if that is what’s happening.