james debate
james debate

Sunday, 20 November 2022

Last week, America went to the polls and elected a new Congress. This blog most recently posted on the subject with our final forecast. Now that the dust has settled (aside from a few straggler House seats and a Georgia runoff) it's time to step back and take stock of what happened. Were we surprised? What will the consequences be over the next few years? Most importantly, what does this all tell us about the direction the country is headed?

2022 us midterm congress election house senate trump democrat democratic victory republican defeat historic shock

Results
It's safe to say that this was not the midterm election the Republican party expected.


As I noted in my midterm election preview, midterm elections tend to heavily favour the opposition party. In that preview I discussed a few of the reasons why this is the case. Whatever the reason, they tend to win, and they tend to win by a lot. The average opposition party gain in a midterm election is some 30 seats in the House and 4 seats in the Senate. In recent elections the gains have been even more severe: Democrats gained more than 40 House seats in 2018, Republicans gained 60 in 2010.

There is such a strong correlation between midterm election success and being in the opposition party that it has almost become a cardinal rule of US electoral theory. In the modern era, the only times this principle has not held have been under exceptional circumstances: 9/11 and the Clinton impeachment.

This year in particular had all the makings of an ideal storm of high inflation, mediocre Presidential approval ratings and narrow incumbent margins. It should have been a bloodbath. For this reason, all the talk in the build up to election night was not so much over who would win, but by how much the Republican party would win. The phrases "red wave" or "red tsunami" were thrown around a lot.

But this is not what happened. The Democrats have held their majority in the Senate and appear set to even increase their majority. The Republicans did take a majority in the House, but by the barest of margins with gains likely to be in the single digits, resulting in potentially the smallest House majority of all time. They went into election night with their majority considered a mere formality. In the end, they are likely to have gained 6-8 seats in the House and lost 1 seat in the Senate.

This really can't be stressed enough. This was a catastrophic election for Republicans. Despite a turnout advantage and a seemingly ideal electoral climate, they barely managed a tie. This was a historic electoral achievement for President Biden and the Democrats, and 2022 will go down in the history books as one of those asterisk elections, the one that years from now people will look back at and ask if the next election could be "another 2022".

So the question now is: how did this happen, why didn't anyone see it coming, and what does it mean going forward?


Why did the GOP underperform?
Analysts will be poring over this one for years, trying to figure out how the Republicans performed so badly. Democratic turnout was as tepid as one would have expected in a midterm election, but independents broke unexpectedly for Democrats. Why?

A key early observation is the fact that the Trumpist "MAGA" candidates all fared very poorly. All across the nation, election deniers and far right extremists lost the close, toss up races. In the big contests, they fared particularly dismally. Nowhere is this more clear than in Arizona. A traditionally red state, Arizona was swept by Democrats, winning a key Senate race as well as the Governor's mansion. On the losing end? MAGA darling Kari Lake, a TV personality who in the run up to the election had been tipped as a future star and potential Vice Presidential candidate for Trump's 2024 run. 

On the other hand, Republicans fared better in states where Trump and his brand of politics were kept at arm's length, most notably in Florida under the banner of Trump rival Governor Ron DeSantis. Indeed, the Sunshine State was one of the few bright spots in an otherwise bleak spate of results for the GOP. Whatever else the data shows, it is clear that those closely associated with Trump fared worse than those that kept their distance.

An interesting datapoint is that Democrats appear to have overperformed most in the competitive battleground states, while their performances in the less competitive states falls more in line with pre-election expectations. This suggests that their campaign messaging has been particularly effective, a stark rejoinder to the commonly held wisdom that Democrats have lagged their rivals in this regard.

Republican Senate leader Mitch McConnell, of all people, summed it up perfectly when he said that Republicans lost the election because voters saw them as causing "too much chaos". I think he's hit the nail on the head here, but it's not the only factor.

A significant factor in the GOP's underperformance that should not be ignored is the redistricting process of the last two years. As I discussed in my pre-election preview, Republicans have had a significant structural advantage in recent midterm elections due to an election rigging process called gerrymandering (see that post for further explanation). Under the old districting, it was estimated that Democrats needed to win nationally by around 5% just to break even in the House. This is essentially what happened in 2020 when they won by that margin and held only a small majority. Under the new districts, that structural advantage has narrowed considerably and may even have vanished entirely. 

But just as other "asterisk" elections tended to be precipitated by significant events (9/11, impeachment), we also need to consider the exceptional occurrences that may have swung this election in Democrats' favour. 

Most notable among this has to be January 6. On January 6 2021, Donald Trump attempted to overthrow a democratically elected US Government by staging a desperate and ultimately deadly insurrection in Washington DC. This may well be the single most traumatic blow to the American psyche since 9/11. America was attacked on this day and for the first time since the Civil War, that attack came from within. For the first time in modern history, the very principle of American democracy was under threat. Hundreds of criminal charges have been filed in connection with that day, including multiple charges of insurrection. It's also been the subject of a major House investigation and now a Federal Special Prosecutor investigation. If anything can be said to be a significant, election swinging event, it's this.

The common wisdom leading up to the election was that the shock of this event had already worn off. That the American public, somehow, wasn't interested or didn't care about this assault on democracy. The media was full of premature hot-takes that Democrats should have spent more time focusing on kitchen sink type issues, that their grander arguments on ideals and principles would not reach the average voter. This has proven to be well wide of the mark. Voters did, of course, care about this horrific and shocking moment in American history, and it appears to have significantly poisoned the Republican brand. A reminder to take media punditry with a pinch of salt, the people running the campaigns usually have better data than they do.

Remarkably, this was not the only such event in the build up to this election. In summer 2022, the Supreme Court voted to overturn the decades-standing precedent that a woman is entitled to bodily autonomy, the case of Roe v Wade. It sparked widespread outrage all over the country, but as with January 6 the common wisdom was that the anger had worn off by election night, that voters were apathetic and unconcerned.  Again, this was clearly wrong. The anger has persisted through to election night and it has cost Republicans big.

As I said, this is a question that will be talked about for years to come. But when all is said and done, the reasons for GOP underperformance in 2022 are likely to be a combination of poor candidate quality, losing their gerrymandered advantage, and the two opposition-rallying events of Jan 6 and Roe v Wade.


Why didn't anyone predict this?
While the GOP's disappointing electoral performance will rightly remain the big story of the 2022 election, an equally worthy story is the fact that all the experts, all the forecasters and analysts did not see it coming. 

The media was unanimous in the build up to the election. This was going to be a red wave, a historic and massive victory for the Republican Party. Fivethirtyeight forecast an almost 90% chance of Republicans winning the House and a 60% chance of winning the Senate. RealClearPolitics predicted they would gain 50 seats in the House and sweep the competitive Senate races for a whopping 4 seat gain. Media all over the world were certain of this outcome regardless of their political persuasion, from the far right propagandists of Fox News to left leaning outlets such as the Guardian. They all missed it.

But not everyone missed it. On this very blog, a week before the election, I predicted that Democrats would hold the Senate at a time when few others were doing so. I also called the House race a toss up at a time when just about everyone else viewed it as a landslide, and predicted a much smaller than expected GOP majority. At the time of writing, I have correctly predicted every single Senate race that has been called. I'm not a professional by any means. So how did I, an admitted amateur, see what the professionals did not? It wasn't hope, it wasn't wishful thinking, it was there in the data, clear as day.

As I noted in my preview, the numbers really weren't that good for Republicans. Democrats were fundraising at a high level, special elections were breaking for Democrats by margins far above their expected baseline. These are very bullish signs for a party's electoral chances. Most importantly though, the polls were not good for Republicans, despite what the experts said.

For most of the summer, polls consistently predicted a close race with perhaps a narrow Democratic advantage, much to everyone's surprise. For the reputable, high quality pollsters, this was the consistent and clear picture coming from the data and it should have been a big red flag for Republicans. Then a month before the election, the polls suddenly swung in Republicans' favour. Analysts interpreted this as some late, unexplained shift in the race. There was just one problem: it was entirely fictional. 

Let's not beat around the bush. There are pollsters out there that are not serious polling firms. They are propagandists whose intent is to try and drive momentum towards their preferred side. They do it every year and yet, somehow, people keep paying attention to them. Get ready, because I will name and shame.

Look at the polls during that period of the race. Look at the actual pollsters. You will see time and time again it was the same few pollsters showing this late movement towards Republicans: Rasmussen, Trafalgar, Co-Efficient and InsiderAdvantage. None of the other pollster showed this movement. None of the gold standard, reputable pollsters saw anything other than consistently favourable Democratic prospects. 

Now look at the 2020 polls. You will see those exact same pollsters predicting a Trump victory. 2018, the exact same thing. These few pollsters do this every election cycle, and it's time we stop taking them at face value. 

Rasmussen is commonly considered as the gold standard of partisan right wing pollsters. One only needs to look at their shamelessly partisan press releases, typically full of Republican party talking points and attacks on the liberal media, to see the type of operation they're running. Their data has always been suspect, not only for its consistent partisan lean, but for its unusual movement and trends that are generally not reflected in other polls. In particular they have been frequently suspected by poll-watchers of pushing right-leaning numbers, and then shifting dramatically to the left with their final polls so that they can point to a respectable final prediction. Ultimately though, the data they produce is just flat out wrong. In 2018 they predicted Republicans to win the popular vote, in a year where they lost by a landslide margin. In 2020 they were equally wide of the mark. In 2022, they predicted a landslide Republican popular vote margin in a year where they are just barely breaking even. To be clear, there is nothing wrong with pollster having a particular "house effect" or partisan lean. So long as the methodology is consistent and honest, they can still provide useful data. But what really sets off red flags with Rasmussen is the fact that they refuse to acknowledge themselves as a partisan affiliated pollster, despite their obvious political lean. If they're dishonest about that, why would anyone trust the numbers they produce? It's time we stop paying attention to these guys.

Potentially even worse than Rasmussen is Trafalgar. Trafalgar has the ignominy of having called every single battleground state wrong in 2022. Even worse, they called several states that were nowhere even close to a battleground, wrong. They predicted Republican victories in Michigan, Colorado, New Hampshire and Washington. They lost those races by double digits. To call Trafalgar's numbers suspicious would be an understatement. Trafalgar consistently come up with polls showing very narrow Republican leads in every race, regardless of state. States that look nothing alike electorally, which end up with results that differ from one another often by double digits, and Trafalgar invariably finds Republicans leading all of them by the same 1% margin? It stinks to high heaven, even more so when the pollster reveals nearly nothing about their methodology. Trafalgar's methodology is secretive, their results are incoherent and make zero sense in context, and their track record is abysmal. They are not a serious pollster and it's time to stop taking them seriously.

These are what I would consider the "big two", but there are others that are worth mentioning. Co-Efficient is a Republican affiliated pollster who made headlines for predicting a Republican victory in New York, one of the bluest states in the country. During October they furiously spammed the polling averages with an absurd volume of polls, all of which boasted similarly outlandish predictions of Republican victory in blue strongholds, none of which panned out. Then there is InsiderAdvantage, a pollster run by noted Republican propagandist and TV host Sean Hannity. Needless to say, his polls were not accurate. That didn't stop them from having an outsized effect on the polling averages though.  

I can criticise these pollsters all day, but for what? Ultimately the key point here is that these were the only pollsters showing this mythical red wave. The reputable pollsters showed nothing of the sort, but their results were overwhelmed by the sheer volume of polls that these right wing organisations churned out to "flood the zone" and manipulate the averages. They were the ones pundits and forecasters were relying on for their predictions. It was obvious that these guys were just trying to game the system, so why did the "experts" fall for it? That's a really good question.

A big part of the problem this year was the lack of high quality polls. For whatever reason, the traditional pollsters were largely MIA, while these right wing pollsters flooded the zone with their propaganda. This meant that the polling averages and electoral forecasts were generally dominated by these junk pollsters. What little high quality polling data we did have painted a very different picture, but for whatever reason the analysts ignored this, and allowed the junk polls to influence their forecasts.

I suspect a big reason for this is all the anti-polling discourse that has become fashionable in recent years. I have written frequently about how this is mostly hype: the polls weren't actually that bad in 2016 or 2020 they were just poorly interpreted, and the polls in 2018 were pretty much dead on. Nevertheless, it appears even the experts have been influenced by that hype, and the 2022 coverage has been characterised by an inherent distrust of the polls. With this in mind, it's not surprising that fewer people were commissioning polls from the mainstream pollsters, and that a disproportionate number of the polls that were commissioned were commissioned for political purposes. 

At a time when we have been (unfairly) conditioned to mistrust polls, it is not surprising that many fell back on fundamentals and historical precedent for the basis of their predictions. The problem with 2022 is that they relied on this too much, while failing to properly evaluate the data that was out there.

A special mention also needs to be made of RealClearPolitics. RCP is, ostensibly, a polling and political news aggregator, non-partisan in nature. That's all a facade though. Their political bias and fickle methodology has been widely criticised in the past. In particular they have a tendency to pick polls that favour Republicans, then ignore those same polls if the results get worse for them, and they arbitrarily set cut off dates for polling averages depending on what gives Republicans the best result. RCP has been pretty bad for years now and, much like Rasmussen, getting worse with every election cycle. 2022 marks a new low for the organisation, with a final forecast that was embarrassingly wide of the mark, predicting Republican victories in states that were never competitive and lost by wide margins. 

If there is one thing we can all take away from the 2022 election, it is that there is no reason whatsoever to listen to anything that comes from RealClearPolitics, Rasmussen, Trafalgar or their ilk ever again. They are not serious organisations and should not be treated as such. It is time we all recognise this.

There is no denying it. 2022 was a very bad year for the pundits and forecasters. They ignored the high quality data (which in fairness was lacking in quantity), allowed themselves to be swayed by trash data. They listened to the polls that supported their preconceptions and expectations and disregarded anything that didn't. The data was there. I pointed it out in my preview. They just ignored it for some reason. 


What does this mean going forward?
There's no sugarcoating it. Republicans are in big trouble. If they couldn't win in the highly favourable electoral climate of 2022, what hope do they have in 2024, a year that will almost certainly be more favourable for Democrats?

The answer is "not much". Not unless something changes. If Republicans go into 2024 still championing the crazies and conspiracy theorists, they will lose big. The question is whether they will learn the right lessons from this setback.

More immediate is the question of how Government will function these next two years. With split control of Congress it appears unlikely that much of anything will get done. The new Republican House majority has indicated that they intend to prioritise political theatre, fake investigations of Democrats and their families. First they will need to actually agree on a House Speaker, however, and with their tiny majority there's very real doubt about whether they will be able to do so. There has even been serious talk of a compromise candidate between Democrats and moderate Republicans. 

Additionally, with such a small majority, can we even be sure they will hold that advantage until the next election? Every single retirement, every death, every defection and special election going forward is going to be a matter of national import, potentially one that could swing Congressional control. We will see what happens, but this could get into some pretty murky, unprecedented waters.

In particular, this election was a disaster for Donald Trump. For the first time since his political ascent, his domination of the Republican Party appears to be in doubt. His allies and chosen candidates all fared miserably, while his political rivals did well. Many elected officials, news organisations (Murdoch in particular) and, indeed, voters, appear to be breaking with him, having finally had enough. Post election polls now have Ron DeSantis as frontrunner for the GOP nomination, with large numbers of Republican voters not wanted Trump to run at all. It will be very interesting to see what happens next. It is clear that many Republicans will no longer back him, but at the same time his diehard basis is very loyal. Whether anyone can find a way to unite this fracturing political party ahead of 2024 will be the big question.

Perhaps the most significant takeaway from the 2022 elections (in terms of electoral politics) will be the apparent end to the GOP's semi-permanent gerrymandered advantage in the House. Between 2010 and 2020 it really was the case that the House was essentially out of reach for Democrats, even in years where they were winning by wide margins. Following redistricting, those days appear to be gone. Democrats have managed to essentially tie the House race despite having little if any national advantage in the polls. It could be a fluke, but right now it appears that their gains in redistricting may have been even greater than initially thought. The result could be that the House elections are much more competitive in the years to come.

Finally, there's the question of the polls and forecasters. We need to be very clear on this. The polls in 2022 were generally quite good, but there was a lot of junk data out there muddying the waters, and the "experts" did a very poor job in critically evaluating between them. Sadly, I fear people will take the wrong message from this election and distrust all polls, rather than discerning specifically those that deserve to be distrusted. We need to do what we can to make the general public more data-literate, and start calling out the con-artists in political media for what they are.

So there it is, 2022 election is done, and the first battle lines have been drawn for 2024. This is an election that will be remembered as a historic accomplishment, as well as a potential turning point in the politics of the nation.








Older Post Home