Demographics aren't destiny
For some time now, there has been a discussion online and in newspapers about Democrats’ declining support from non-white voters. It was turbocharged over the week after a thread from a data analyst at the Financial Times charting and graphing the alleged trends was posted on X. A recent poll from the New York Times, which is cited in the Financial Times thread, showed Biden holding steady with whites, but bleeding support from non-whites. Nate Silver wrote a piece about Democrats losing support from non-white voters in response to the thread and poll. David Leonhardt in the New York Times has had two pieces out discussing the rightward shift of non-white voters and why it is happening.
I find the whole discussion to be noteworthy, but also a bit odd. It comes at a time when Democrats have been winning elections across the country. Right now, Democrats have a majority in the Senate, almost a majority in the House and are in better shape on the state level than they have been in at least 20 years. Yet much of the discourse centers around their coalition supposedly collapsing. I guess it’s true their coalition is falling apart… everywhere except the voting booth.
The freakout (or celebration) among the chattering classes about Democrats’ losing support from non-white voters reminds me of the discussion going on a decade ago that was just the opposite scenario. After Obama was reelected, it was conventional wisdom among many pundits and others that he had brought together the “coalition of the ascendant.” The country was becoming more diverse and he had successfully united non-white voters of all groups. In addition, younger whites were more Democratic than their elders and would soon overwhelm them.
Republicans were in deep trouble. They were an almost all white party in an increasingly diverse country. Their coalition was old and dying off. They had to change their ways fast or they would risk becoming a permanent minority.
That’s not how things panned out. It was assumed Republicans would never win the White House again unless they supported comprehensive immigration reform. They nominated Trump in 2016 and he won despite being rabidly anti-immigration. He lost in 2020, but made big improvements with Hispanic voters.[i]
I think it’s best to see the developments from Obama up to now as offshoots of bigger forces at play. In the case of Obama, he had assembled a very diverse coalition and did very well among young and non-white voters. That’s only part of the story though. He also did well with non-college educated white voters, especially in the midwest. That was something the “coalition of the ascendant” story tellers conveniently overlooked.
Obama was an extraordinarily good candidate. He had a strong personal appeal with many voters that was unique to him and didn’t carry over to Clinton or Biden. He also benefitted from things beyond his control. When he first ran in 2008, Republicans had been in the White House for eight years. That counts against the candidate from the party in the White House. Making his job even easier was the fact that Bush was very unpopular. On top of that, the economy fell off a cliff, which all but sealed his victory. In 2012, he got to run for reelection against a weak challenger who was easily painted as a plutocrat.
The limits of the Obama coalition first became clear in 2010. Democrats were crushed nationally and on the state level. Many of those who voted for him in 2008 were nowhere to be found. Most of them came back in 2012 and he was reelected, but in 2014, they were nowhere to be found and Democrats got crushed again.
After the 2014 midterms, Democrats were in the worst shape they had been in on the state level since the 1920s. I don’t blame Obama for that. His job was to get elected and reelected president and he did it. Candidates do what they need to do to win and every candidate will assemble their own coalition. Because so much of the party infrastructure didn’t support him in the primary in 2008, he had to build his own organization. It worked out great when he was on the ballot, but it was terrible when he wasn’t.
Republicans are now running into many of the same problems Democrats ran into during the 2010s. Trump has a strong personal appeal with many people, but it’s unique to him and doesn’t carry over to anyone else.[ii] Many of Trump’s avid supporters are voting for him, not the Republican Party. Since Trump won, Republicans have done poorly almost every year.
Like Obama’s coalition, Trump’s coalition is reliant on low propensity voters. When he’s not on the ballot, many of those voters don’t show up and Republicans suffer. The key ingredient is education polarization. Those who are college educated are much more likely to vote regularly than those who are not. During the 2010s, college educated voters often favored Republicans, especially college educated whites. Since the GOP became the party of Trump, college educated whites have flocked to Democrats.
Looking at the increasing support from non-whites for Republicans, it becomes clear, as Leonhardt notes, that it’s largely a story of growing education polarization. With non-college educated voters becoming more Republican, that has meant more non-white voters have become Republicans. Conversely, with college educated voters becoming more Democratic, that has meant more white voters have become Democrats.
It also has to do with ideology. There are many non-white voters who are conservative who have voted for Democrats because of identity or tradition. That has now begun to erode and more conservative non-whites are voting Republican. That has been especially pronounced among Hispanics.
What, me worry?
In a reversal of the discussion from a decade ago, it’s now Democrats who are doomed. Republicans have been making gains among non-white voters very fast. At the same time, non-college educated white voters, already a Republican constituency, are becoming even more Republican. The gains Democrats are making among college educated white voters are too small to offset that and the number of people attending college is down so in the future their numbers will be even smaller. With those demographic trends, Democrats are in danger of being in the minority.
The discussion a decade ago was wrong and the discussion now is almost certainly wrong, too. I’m not a believer in demographics being determined. It assumes things are fixed and will remain constant, which is the antithesis of my thinking on most everything. Trying to predict the near future is hard. Trying to predict the future a decade from now or later is all but impossible.
Demographics do matter, but not because they will make one party dominant nationally. In individual districts and states, demographic trends can be very consequential. There are parts of the country that were solidly Republican until Trump showed up and now are either swing districts/states or have become reliably Democratic. There are parts of the country where the opposite has happened.[iii]
For the country as a whole, not a lot has changed. We still have two major parties. Both are capable of winning the White House and Congress. What looks like a significant advantage one party has is almost always as an artifact of a particular cycle(s) and/or a candidate being unusually strong with key groups.
For example, one big myth that thankfully is now dead is the “blue wall.” That mostly referred to Michigan, Wisconsin and Pennsylvania having voted for every Democratic presidential candidate since 1992. The assumption was those states would keep doing that.
Gore and Kerry both won those states, but very narrowly. Obama did unusually well there and gave some the illusion that those states were solidly in the Democratic camp. Like I said, he had unique strengths that didn’t carry over to Clinton or Biden, which is why the former lost those states and the latter barely won them.
Since 2020, another myth has emerged. That is the idea that Republicans have a durable advantage in the electoral college. Trump did much better in swing states than he did in the national popular vote. Despite losing by 7 million votes (and 4.5 points) in 2020, a 40,000 vote shift in Arizona, Georgia and Wisconsin would have won him another term.
As the nominee this year, many in the chattering classes think he will have a similar advantage. If anything, his advantage may be even bigger. Biden, the thinking goes, will have to win the popular vote by even more just to have a 50-50 chance of getting reelected. Anything is possible, but that is hardly set in stone. In fact, there is a strong case to make that Trump’s electoral college advantage is narrowing and may disappear. Even if Trump maintains an electoral college advantage, it probably will be because of unique strengths he has with key groups and not something that carries over to other Republicans.
As for the rightward shift among non-white voters, a big question is how consequential it could be. We don’t know yet how enduring it is. We do know it happened in some places in 2020 and 2022, but whether it’s a lasting trend or a blip is to be determined. We will probably need a few more cycles before we can have a firm idea.
It’s important to remember that these shifts are relative. Non-white voters have moved towards Republicans in recent cycles, but they still heavily support Democrats. See the chart below.
Looking at the changes from 2016 to 2020, Biden did a good bit worse with Hispanics than Clinton, but he still won a solid majority. Unlike Clinton, he won the election.[iv] I like to use Arizona as an example of how the shift in non-white voter preferences can be overthought if looked at in a vacuum.
Arizona has a large Hispanic population, but was a reliably red state until recently. Biden won it and Democrats did well there in 2022. I haven’t dug deep enough to know for sure, but it could be that the shift among Hispanics towards Republican is much smaller there while the shift towards Democrats among whites is fairly significant. Conversely, if the shift among Hispanics towards Republicans was big, then the shift among whites towards Democrats was even bigger. Either way, Democrats won there in 2020 and 2022, which is what counts.
To give another example of why demographic trends don’t worry me, take a look at the graph below. It’s from Leonhardt’s first piece I linked to. It shows how Democrats did with non-college educated voters as a group and by race in the 2014 and 2022 midterms. As you can see, Democrats did worse in 2022 with each group, albeit not by a lot.
Why am I not worried when looking at that graph? Because Democrats did horribly in 2014 and kicked ass in 2022. Despite doing worse with a big part of their coalition, Democrats had the best midterm for the party in the White House since 1934. Presumably, they did much better with college educated voters and that is the key in non-presidential elections.
The main reason I don’t worry about Republicans gaining a durable advantage by doing better with non-white voters is the same reason those voters didn’t give Democrats a durable advantage. This is where education level really comes into play. Non-white voters who are college educated basically support Democrats as much as they did before Trump came along. It’s non-white voters without a college degree who are shifting towards Republicans. That’s problematic for Republicans because those are low propensity voters. Most of them vote only in presidential elections.
By becoming the party of low propensity voters, Republicans are now at a disadvantage in non-presidential elections just as Democrats were last decade. If the trade going forward is Republicans gain a modest advantage in presidential elections while Democrats gain a big advantage in all other elections, sign me up right now. Of course, presidential elections matter, especially this one, but so do all other elections. Control of Congress, states and localities is frequently determined in midterm and other kinds of elections. Very few offices hold their elections exclusively in presidential election years.
Macro features to keep in mind
The idea of non-white voters shifting towards Republicans is surprising psychologically more than anything else. The dominant narrative among many of those in the chattering classes, at least until recently, is that non-white voters are almost all Democrats and will remain that way so anything going against that shatters that narrative. It shouldn’t be surprising that it’s happening when taking into account some of the country’s big features.
We have two major parties and always have. Demographic changes have been happening for over 200 years, but two major parties have always been dominant. We have had the same two since 1860.
Decades from now, the country will be much more racially diverse than it is today. We will still have two major parties, almost certainly the same two we currently have. Inevitably, that will mean lots of non-white people voting Republican.
There have been periods where one party enjoyed decades of national dominance, but that was long ago. Since 1968, neither party has had an advantage. My guess is that trend will continue well into the future.
There’s a very high amount of partisanship and polarization compared to the last time one party was dominant. Very much related to that is the media atmosphere. Gone are the days when Edward R Murrow and Walter Cronkite were the voices everyone knew. The media ecosystem today is highly fragmented. There is no single source that most people rely on.
Because of media fragmentation, people can hear what they want to hear. With the internet and now social media, everything is national. Tip O’Neill’s famous quote is way out of date. People now hear about trivial things that are happening thousands of miles away from where they live and it shapes how they vote.
Because of heavy partisanship and polarization, tolerance for bad behavior from one’s own side is much higher than it used to be. For those wondering why Trump’s behavior doesn’t end his chances of winning, you can thank those two things. This day and age, most people will stick with their side no matter what. Even repulsive candidates stand a good chance of winning. Trump is the most famous, but he’s not the only one, i.e., Kari Lake.
Heavy partisanship and polarization aren’t going away any time soon. There is nothing any candidate or party can do to dislodge it. If it does end, it will be because of changes in norms and culture, not changes in policy and definitely not because a white knight candidate came down from the heavens and brought everyone together.
Both parties today are guaranteed a high floor and low ceiling of support. The days of landslide victories like Nixon in 1972 are over. Anyone who is the nominee of a major party for president has a much greater than 0% chance of winning no matter who they are.
It’s also important to remember that parties do adjust to changes. If it looks like one party is at risk of being at a disadvantage, they will change in some way(s) to prevent it from happening or continuing. If the prophecies of doom currently being made about Democrats look like they might come true, they will adjust. Whatever it takes to start winning enough voters to win the White House and Congress, they will do it.
How do parties adjust? Usually by waiving the white flag on certain issues. That’s what Trump did in 2016. The Republican position on Medicare and Social Security was toxic and he disavowed it. That had the effect of making those programs a non-issue in the election and disarmed Democrats of an effective weapon.
If Democrats lose this year, they may not change because it usually takes two presidential election losses in a row. If they continue to lose they will change on whatever issue(s) are causing them the most problems even if it upsets some in the party. Parties don’t like to fight among themselves, but they hate losing even more.
When looking at all those factors (persistence of the two parties, partisanship and polarization, media fragmentation and parties making adjustments), it’s best to think of demographic trends as being like the stock market. That is to say they move randomly and unpredictably over short and long time periods. Trying to predict where the stock market will be in 10 years is quixotic. The same is true for trying to predict what will be the dominant demographic trends in the 2030s.
The demographic trends happening now may turn out to not have much of an impact when all is said and done. In the future, there may be demographic trends that have a major impact that nobody is thinking about today because they haven’t happened yet. What could those be? Your guess is as good as mine.
[i] It was a shock that he did was well with Hispanic voters as he did only because he’s Trump. It’s not unheard of for a Republican to do well with Hispanic voters. George W Bush did quite well with them in 2004.
[ii] What makes it worse for Republicans is Trump’s appeal is much narrower than Obama’s was. Obama’s share of the vote was 53% and 51% while Trump’s was 46% and 47%. When Obama was on the ballot, Democrats did well almost everywhere. When Trump was on the ballot, Republicans did well in some places, but not others.
[iii] Colorado, Virginia, Florida, Ohio and Iowa were swing states until recently. North Carolina, Georgia and Arizona were reliably red states not long ago. Texas isn’t a swing state yet but that may change soon. Education polarization is the main driving force behind all that, for better or worse for Democrats and Republicans.
[iv] No joke, from much of the media coverage of the aftermath of the 2020 election, you would have thought Trump won. A whole lot of it centered around Trump doing better with Hispanics, never mind that Biden won Arizona and Nevada. Coverage of Texas was all about the Rio Grande Valley moving sharply to the right. You would never know that Biden did better in Texas than any Democrat had in over 20 years. That’s because he made big inroads in places where people actually live like Austin and DFW.