October 15: Biden Is the Early Favorite For Reelection Even If Dems Lost Ground in 2022
After absorbing a lot of Democratic gloom-and-doom about the midterms, I offered some silver lining at New York:
The 2022 midterms don’t look great for Democrats, who will try to buck history by hanging on to super-slim congressional majorities. Thanks to the particular lay of the land, Democrats have a decent chance of maintaining control of the Senate. But the House? Not so much: The two times since the New Deal when the president’s party won net House seats in a midterm (1998 and 2002), the president in question had sky-high job-approval ratings. Even if you believe Joe Biden’s plunge in popularity has been stemmed or even turned around a bit, he’s not going to have 60 percent-plus approval in November 2022 unless really crazy things happen. There’s just too much partisan polarization for that these days.
Thankfully for Democrats, even if they lose their congressional majorities next year, Biden himself won’t be an underdog for reelection in 2024. After all, the last two Democratic presidents were reelected after historically terrible midterms. Democrats lost 54 U.S. House seats in 1994 and 63 in 2010. Yes, they had bigger majorities going into those elections than Democrats have now. But they lost the national House popular vote by an identical 6.8 percent in both midterms, which is pretty bad, particularly since Democrats suffer from a voter-inefficiency problem in House elections (too many voters concentrated in too few districts).
It’s possible for a president’s party to lose a midterm so badly that bouncing back in the next cycle is all but impossible. Consider the man whose unique comeback accomplishment Donald Trump will be emulating if he runs in 2024, Grover Cleveland. The president Cleveland defeated in an 1892 rematch, Benjamin Harrison, was a Republican whose party lost an incredible 93 House seats in the 1890 midterms. This, mind you, was at a time when the House had only 332 members, which means the GOP lost over half their caucus in one cycle (an even worse percentage than in 1894, when Democrats lost a record 125 House seats during the midterm after Cleveland’s comeback triumph). In this era of polarization, nothing like that is going to happen to Democrats in 2022.
Looking more broadly at the power of incumbency, there have been 13 sitting presidents since World War II who were on the general election ballot. Nine of them won. The four losers all faced special circumstances. Gerald Ford had not previously been elected to anything more than the U.S. House; he ascended to the vice-presidency and then the presidency when disgraced predecessors resigned, and he pardoned the president who appointed him, the especially disgraced Richard Nixon. Jimmy Carter was caught up in a historical realignment that he had held off four years earlier by carrying his native South, which then resumed a massive Republican trend. George H.W. Bush suffered from a terrible economy but then also a party split (third-party candidate Ross Perot won a lot of previously Republican voters). And we all know about Donald J. Trump, who was impeached twice and seemed determined to offend swing voters.
In retrospect, what’s most remarkable is that Ford and Trump very nearly got reelected despite their handicaps, exhibiting not the weakness but the strength of incumbency. And it’s with that perspective that any early handicapping of a potential 2024 rematch should be considered. Trump benefited from incumbency in 2020, as will Biden in 2024. So the idea that the 45th president has some built-in advantage over the 46th — absent the renewed election coup so many of us fear — doesn’t make a lot of sense.
A few comments:
1. That Hawaii poll does not indicate a Bush win there. Far from it. The incumbent is at 43% — not exactly a ‘re-elect’ number. Kerry will surely win Hawaii by 10-15 points or more. Gore won it by 17.
2. There is no “sudden surge” by Bush in Michigan. One poll comes out and shows a narrowing (going against every other poll showing Kerry has indeed captured the state) and suddenly it’s panic time. You have to take all the polls together, and then with a degree of skepticism. Nobody truly thinks Bush has a chance in Michigan. (Ohio also clearly is in the Kerry column.)
3. As Ruy and numerous other pollsters continue to say, in the home stretch the only important number is the INCUMBENT’s number, not the spread. Even in these swing states where Bush supposedly has pulled ahead, in almost no cases does he have 50% of the vote — in fact, in most he is peaking out at 47% or lower, which means that is likely the MOST he’ll get in those states. Not enough to win.
4. The same holds true nationally — sure, the Zogby poll shows a 47-45 Bush lead. But 47 is as high as Bush ever gets in that poll, and it ain’t enough! The undecideds will overwhelmingly break toward Kerry.
5. I honestly believe that given the undercounting of Kerry voters (due to the ‘cellphone-only’ factor and the fact that none of the 30 MILLION NEWLY REGISTERED VOTERS are counted in any poll), that if Bush is not ahead by 5 points in any state he will lose that state. That is not even wishful thinking — that is reality.
You heard it here. Now everybody relax and go put out a yard sign!
Markydeee
College student registration was settled by the supremes in 1979. College students have the option to vote in their home town OR choose to register in their college precinct to vote. This is no longer in doubt. Absent a reopening of the issue by Congress it will remain the law of the land. Local officials may try to suppress College student votes but the efforts will ulltimately fail just by showing election officials the text of the ruling.
Ultimately there will needto be a model voter registration standnard established. In the interest of democracy and access I hope that the MN and WI same day registration laws are used for the National model. Both states have used these approaches for 30 years with very low rates of electoral irregularity.
The MN Republican Sec of State has complained about same day registration citing potential fraud to other states but there is absolutely no desire to touch the statute by either party in MN as there is a common agreement that ballot access is important for the health of democracy.
Very odd this is but the ‘Honolulu Advertiser’ has Hawaii dead even at 43% each with 10% undecided. Friends in Hawaii say it’s highly unlikely that the state goes for Bush but a wakeup call for state Kerry people nonetheless
Zogby is fretting, I think, because he went on a limb early by predicting it’s Kerry’s race to lose. As a pollster with a solid reputation (if I remember correctly, he was not, however, as accurate in 2002 as he was in 2000) he has to be super-sensitive to all indicators to know when to change his mind (I hope the answer is never) in order to preserve his credibility. I also suspect that he is wondering whether he could rely on the last-minute undecideds trend this year since there is such a noisy fear factor. TIPP also went today from +1 to +4, but then Rasmussen went from +3 to +1, and WP/ABC stayed the same, so there is, knock of wood, no sign of a consistent trend. We need to be prepared for a wider spread in Zogby tomorrow since yesterday’s totals were +3 for Bush. Unless Kerry has a very strong day today, Zogby may actually go to +3. But then last week it got to +4 only to come back to a tie several days later. No reason to fret. Plus Zogby is the first one to say that if we have a record turnout, Kerry most likely wins, regardless of how the horserace looks right now.
Comparing the new Time and Newsweek polls provides stark proof that there are systematic errors in much of the polling. If you adjust both to a 38/34 party split (Time gives you their exact split and Newsweek gives you enough data to estimate it within a point), you find that Bush is ahead among RVs by slightly over 5% in the Time poll and Kerry is ahead by slightly over 6% in the Newsweek poll. Since fixing the party ratio cuts the sampling error (the 95% confidence level with 1000 respondents goes from 3.1% to around 2% or 2.1%), these results differ by around five standard deviations. The chances of this happening by chance are on the order of one in a hundred thousand. Clearly, one or both of these polls has a big systematic error in their RV data. Let me emphasize that this has nothing to do with likely voter models.
Also, here’s something new on the Wash. Post tracking poll methodology page:
“The Post also adjusts the percentages of self-identified Democrats and Republicans by partially weighting to bring the percentages of those groups to within three percentage points of their proportion of the electorate, as measured by national exit polls of voters in the last three presidential elections.”
HOWEVER: The Post poll, adjusted to a 34-38 R/D split, still tends to look pretty good for Bush. And adjusting for the R/D ratio will largely (I think) correct for the Post’s failure to weight for Hispanic origin. As I have mentioned elsewhere, the ABC method of weighting (by “cell” rather than by “parameter”) avoids a lot of the pitfalls that other polls run into — it’s not clear whether the Post uses the same method.
The key issue is: Why does Post/ABC and Time show independents favoring Bush by 5 points or so, while other polls show them strongly for Kerry? There is too much of a pattern here for this to be random, and the cause must lie in some aspect of the polling or the analysis. One or the other group of polls is simply wrong (and I can’t exclude the possibility that both are wrong).
Can someone please tell me about the polls in Michigan?
I thought that state was solid Kerry, but now only leaning Kerry?
http://www.nytimes.com/2004/10/24/politics/campaign/24c…
“The Gore states in play are Iowa, Michigan, Minnesota, New Mexico, Pennsylvania and Wisconsin. Of those, analysts and aides said Mr. Bush had the best chance of winning Wisconsin, Iowa and New Mexico. A sudden surge by Mr. Bush in Michigan, a state that Mr. Kerry thought he had put away, caught both sides by surprise, and both men scheduled last-minute trips there for next week.”
Zogby’s remark regarding an “ominous” sign for Kerry was the fact that his polling showed Kerry and Bush roughly equal among seniors. This is consistent with Democracy Corp’s most recent findings, and is difficult to understand considering the issues of prescription drug benefits, Medicare increases, and threats to the Social Security system. I, too, reacted with some alarm at Zogby’s evalaution. But it helps to remember that his polls have a 2.9% margin of error, so nothing has statistically changed. This election, more than any other in recent history, is about GOTV! Especially significant will be first-time voters, particularly between the ages of 18 and 25. They may well mark the difference between success and failure. Also, if record-breaking numbers of Americans vote, even currently out-of-reach states like Arizona (my home) may come into play on November 2nd.
There are two conversations which seem to go on nonstop here.
The first is about the polls, and what can be read into them by virtue of their data.
The second is about how the poll spinning is being used by some in media to help Bush.
I’ve no doubt the second has been established, and will continue.
The first, however, remains in issue. ALL of the polls are undervaluing Kerry. ALL of them. Zogby, Rasmussen, Pew – ALL of them.
There is going to be an increase of at least 10 million voters this time, and they are not voting for Bush.
What we have this year is THE EMPEROR’S NEW CLOTHES taken to the extreme by modern mass media. Every day I hear talking heads on CNN say things that they would know cannot be true if they simply read the internals of the polls they use as their Bibles. Wolf Blitzer treats their poll like it was the original Ten Commandments brought down by Moses.
The magpies of media are all clacking “Bush ahead, Bush ahead,” and will continue doing so.
WRONG.
Kerry is way ahead right now. WAY ahead.
November 2nd will prove me right, too.
Let’s just hope that College Students who properly registered are allowed to vote this year. This is an argument Congress needs to settle once and for all, namely is a student resident at a college an actual resident of the local community for electorial purposes? I believe the motor voter law did attempt to set a common standard for length of time prior to an election necessary for legal residence — but many states and local jurisdictions still attempt to apply much more restrictive rules. I know of two cases this year where ther federal standard has been challenged — Prairie View College in Texas (one of the Historic Black Colleges) and Skidmore in upper NY State.
SRBI has yanked the Time poll off their site. Makes it hard to check internals.
I’d like some discussion about the Zogby polls this past week. It’s not so much the poll numbers — Kerry and Bush have remained basically tied all week — but Zogby’s brief comments each day. Yesterday it was that some number or other was “ominous” for Kerry (strong word) and today he muses about whether Independents are breaking for Bush, despite apparently little movement in the numbers. Doesn’t he seem to be exaggerating the import of some of these day-to-day results? I’d like to hear some other comments.
I have it on good authority that several large newspapers that have not endorsed a Democrat for president in decades will do so tomorrow, including the Columbus Dispatch and the Daily Press in Hampton Roads, Va. I realize that endorsements in and of themselves have little impact, but the news media have begun covering the story about the growing number of newspapers that endorsed Bush in 2000 are rejecting him this year. And even more astounding: The American Conservative magazine endorsed Kerry.
The tide has turned.
I see your Time poll, and call with the Rasmussen, Zogby and Newsweek poll which show that it is a tie as of this morning – I have not averaged but if I had to guess- probably at or around 47 each. It remains as it has been- a tie w/ the occasional poll s howing an outlier- anyone on the left or right expecting a break out is living in a fantasy- this is a bruiser, not a TKO type of an election.
On the other hand, Newsweek looks much better than last week (46%-46%) so it and Time just switched places. The horserace is still probably a wash and GOTV in the battleground states is all that matters.
This is sort of depressing since Zogby does not make statements like that easily: Pollster John Zogby: “Bush had a stronger single day of polling, leading Kerry 49% to 46%. For the first time, in the one-day sample Bush had a positive re-elect, 49% to the 48% who feel it’s time for someone new. Also in the one-day sample, Undecideds were only 4%. Could Undecideds be breaking for Bush?” That means tomorrow’s numbers will probably show a larger gap. I think we still have the turnout on our side (unless the GOP thugs manage to play havoc with thousands of ballots) but the undecideds rule is definitely getting pressure from the fear rule — fear is irrational, and people are afraid of change even if things are really bad. Let’s hope it’s not so.
It’s a national poll, so it doesn’t matter. Bush has such a big lead in the South that, of course, the national polls may show a slight lead for him.
The only polls that count are the ones in the Battleground States.
fyi… the latest Time poll that puts W up by 5 has this in the method section:
“Likely voters reported party identifications are: 35% Democrat, 35% Republican, 23% Independents. Registered voters party affiliations are: 35% Democrat, 33% Republican, 23% Independent.”
I’ll let the stat-heads in here tell us if this is skewed..
eric…
post election situation: to dispense with the urban – rural dichotomy that seems to be represented by red v blue, we would need to re-structure electoral college.
waht would the colorado paln nationwide do to electoral collge polling?
“And even more astounding: The American Conservative magazine endorsed Kerry.”
There were several columns, each making a case for their preferred candiCATEGORY: Ruy Teixeira’s Donkey Rising
But the editors couldn’t get behind any single candidate.