Is Sanders or Warren Holding the Other Back?

One line of thinking about the Democratic primary follows a peculiar logic. It goes like this: if Sen. Bernie Sanders or Sen. Elizabeth Warren drops out, you can just add the polling numbers of one to the other. They’re splitting the progressive voting bloc! Either would be winning against former V.P. Joe Biden if the other just dropped out! Right? Except this isn’t the way voting blocs and demographics work.

Voting Blocs Intersect

Yes, there is a progressive voting bloc that favors Sanders and Warren. However, the argument above assumes that voting bloc isn’t intersected by a variety of other voting blocs. Yet it does intersect with both demographic and issue-driven voting blocs:

There’s a voting bloc of white men that favors Biden and Sanders. There’s a voting bloc of white women that favors Warren. There’s a voting bloc of Black voters that favors Biden. There’s a voting bloc of LGBTQ voters that favors Warren. Hispanic voters are generally split fairly evenly between the three top candidates, with Biden and Sanders enjoying a very slight advantage.

People tend to be very comfortable with the perception that there’s a Black voting bloc that’s likely to vote for Black candidates, that women are more likely to vote for women, etc. What’s not discussed or acknowledged nearly as often is that there’s also a white male voting bloc that tends to vote for white, male candidates.

That voting bloc in the 2020 primary largely favors Biden and Sanders. Go back to 2016, and the only demographic split of race and gender that Sanders won in the primary was white men.

Let’s look at second-choice polls in 2020 to see how voting blocs tend to split. Second-choice polls describe who respondents’ second choices in an election are. They’re grouped by first-choice. They do have a margin for error. They’re just like any other poll. However, the fragmentation that’s displayed in them is consistent in nature. Let’s see what happens if Sanders bowed out, and then Warren (just going alphabetically here).

If Sen. Sanders Left the Primary

If Sanders bowed out, Morning Consult’s second-choice poll suggests his voters would split more or less evenly between Warren (31%) and Biden (26%). On the surface, that might not make sense. This is because Sanders’s progressive voters would gravitate to Warren, but white male voters make up the largest segment of his bloc – many would gravitate to Biden. Those who are both would split. There’s a slight edge to Warren, but not markedly so – not to the extent that it would boost her much more than it would boost Biden.

Not all progressive voters are alike. Some would be comfortable shifting their vote to a woman, some less so. Some would be comfortable shifting their vote to a more moderate candidate, some less so.

If Sen. Warren Left the Primary

What if Warren bowed out? A decent amount of her voters go to Buttigieg, possibly because Warren is first by a solid margin among LGBTQ voters. Sanders would get the highest share (30%), but Biden (19%) and Buttigieg (15%) would be close behind.

Sanders’s minor advantage as a second-choice (30-to-19 over Biden, vs. Warren’s 31-to-26) isn’t created by having more appeal. Note that she picks up 31% of Sanders’s voters if he leaves; he picks up 30% of her voters if she leaves. The difference is statistically meaningless. Sanders’s advantage in this poll is that her voters are more likely to chiefly split to three candidates (including him) instead of just two.

Minimizing this advantage isn’t a selective reading. If we use the Quinnipiac instead of Morning Consult second-choice poll, Warren holds a much bigger advantage over Sanders. Here, if Sanders dropped out, she’d be the second choice 35-to-17% over Biden. Whereas if Warren dropped out, Sanders would be second choice 33-to-21% over Biden.

Neither demonstrates a real advantage for one or the other. That’s because the difference is pretty minimal by the time we’re splitting up small numbers into even smaller numbers.

Why the Difference is Slight

By the time you’re splitting up 15-20% of national polling into thirds, the difference between getting 20 and 30% of those voters is a matter of closing a losing margin by one or two percent.

Think about it this way: Biden is polling at 27%, Warren at 16% at the time of this article. Let’s say Sanders drops out. She’s down by a margin of 11% in the polls, but now Sanders’s 20% in national polls is up for grabs.

Warren gains 31% of Sanders’s 20% of voters. That translates to an additional 6.2% in national polls for Warren.

Biden would gain 26% of Sanders’s 20% of voters. That translates to an additional 5.2% in national polls for Biden.

In other words, Biden improves by about 5 points to 32%, and Warren improves by about 6 points to 22%. She’s gone from being down 11 to down 10. That helps, but what we’re really talking about is closing the margin by a whopping 1% in polling averages.

Best Case Scenarios

The very best case scenario for Sanders or Warren getting near Biden would be Warren exiting under the Morning Consult poll. Her current polling average is 16%. Sanders would gain 4.8% in the polls. Biden would gain 3% in the polls. Buttigieg would gain 2.4% in the polls. In other words, Warren dropping out means Sanders would have closed his margin by an astonishing 1.8% in polling averages. Biden’s advantage over Sanders would have started at 27-to-20, and ended at about 30-to-25. I’m sure Sanders would gladly take being down about 5 instead of down 7, but he’d still be down to Biden in the polls.

Let’s take the absolute best case scenario for shifting total number of votes, from the Quinnipiac poll where Warren gets 35% of Sanders’s voters, and Biden gets 17%. That still nets Warren an additional 7%, and Biden an additional 3.4%. She closes the margin by 3.6%. She goes from being down 27-to-16 to being down by a little over 30-to-23.

The point is that neither Sanders nor Warren exiting would put the other over the top. It would come nowhere close. They’d both still be down at the end of the day.

Both historical data and polling data has heavily suggested for the last year of this primary (this has been going on a long time already, hasn’t it?) that one or the other of Sanders and Warren bowing out doesn’t really give the other any major boost.

Voting Blocs Do Not Order by Hierarchy

These are the kind of situations that are likely. Neither Sanders’s nor Warren’s entire number would transfer over. Those voters would split according to a range of priorities and tendencies. You’re not talking about adding Sanders’s 20% in polls to Warren’s 16%. That’s not the way it works. You’re talking about splitting up their voters according to a range of intersecting voting blocs and priorities.

Neither Sanders nor Warren is particularly hamstringing the other. Progressive voters act as a bloc, but that bloc doesn’t exist in a vacuum. White voters, Black voters, LGBTQ voters: all of them have their tendencies and priorities, and those blocs are just as important and functional as a progressive voting bloc is. That’s just dealing with demographic blocs, before even getting to voting blocs that favor single issues.

Looking at politics in a monolithic way, where your voting bloc must magically supersede all others and make them irrelevant and inapplicable, will only teach you to look at electoral histories and polling information inaccurately. Acknowledge and recognize that other demographics exists, and that their priorities can be just as legitimate. When talking about voting blocs, don’t assume they’ll all follow a certain action. They intersect with other priorities that mix their tendencies. Don’t expect other demographic or voting blocs to follow yours. Understand why they might not.

In no world would Sanders’s and Warren’s entire voting blocs combine, even if they were hunky-dory and the best of friends. The best you might hope for in a field with a number of candidates like this is a breakdown where you might get about 40% of a dropped candidate’s voters, and another major candidate gets 20% of them. That closes margins of loss, but it doesn’t come close to evening them out – let alone changing the leader.

Neither Sanders nor Warren is holding the other back from swallowing up the other’s entire group of supporters. Those supporters would split pretty widely and somewhat evenly across the other major candidates because more voting blocs exist than “progressive” alone.

If you enjoy what you read on this site, subscribe to Gabriel Valdez’s Patreon. It helps with the time and resources to continue writing articles like this one.

Why the Same Numbers Mean “Winning” Voters of Color, and “Undecided” White Voters

The narrative goes that Joe Biden’s secured the support of voters of color. Except – wait! The new narrative is that Bernie Sanders might be leading among voters of color. What does this mean? Usually it means that someone is “winning” us because they poll at 20-some percent among nonwhite primary voters, and that’s slightly more than the next person who polls at 20-some percent.

White voters have polled at similar numbers for candidates like Biden and Elizabeth Warren. Why aren’t they “secured” or “wrapped up” or any of the other descriptions we throw around for voters of color?

Biden has the support of a plurality of Black voters in polls. That means more support than anyone else, but not more than 50-percent. That’s often been misrepresented as a majority. In turn, that transforms into a majority of people of color. People don’t bother to look up how Latinx voters, or Asian American and Pacific Islander (AAPI) voters are polled.

Biden’s squarely hung around the mid-20s in terms of Latinx support. With a split field, that’s meant that he still leads by a few percentage points in most polls. Sanders also polls strongly. Warren does in some polls, and hovers nearer 10% in others. Mid-20s is absolutely not a majority. The leader in Latinx polling is often still Undecided.

Add to this that there are barely any polls taken of Latinx voters in the first place, and nobody really has much of a clue – indicators are all over the place and the aggregate itself is shaky.

Websites like Common Dreams and Truthout have run with the notion that Bernie Sanders has just taken the lead among voters of color. Why? Because of a single NPR/PBS NewsHour/Marist poll that places Sanders at 29% and Biden at 26% among nonwhite voters.

That poll holds value, but it doesn’t magically upend all other polls that also hold value. Most importantly, NPR points out about their own poll, “the margins of error with these subgroups are too high to draw any definitive conclusions”. Yet that’s not stopping Common Dreams and Truthout from drawing definitive conclusions.

This is somewhat galling given how insistent Sanders was that voters of color were prioritized too highly in the 2016 primary. Hillary Clinton won 78% of Black voters and 65% of Latinx voters in that primary – a record margin in a primary among people of color – and this was often downplayed as unimportant or only having to do with loyalty.

The idea that anyone – Biden, Sanders, or anyone else – should claim notable support from voters of color because they poll at 20-something percent in an election that has yet to take place is insulting when voters of color voted for a candidate at 70+ percent in the last primary, and this was made to seem thoroughly unimportant.

It makes us into props, easily set aside when inconvenient, but valuable when we help to set the scene for the narrative you want to tell.

Biden and Sanders seem to be ahead of the field among voters of color, sure, but let’s not award them some sort of mandate for polling in the 20s. And let’s certainly not do so when a record margin of support that was 50 points higher has been downplayed and ignored since then by both those candidates.

There are several things going wrong in this thinking. The first is that polls of Black voters represent polls of all voters of color. This is not true. Black voters and Latinx voters sometimes vote in concordance and sometimes support different candidates. That may happen this time.

Terms like “people of color” and “voters of color” have a precise use. They include Black people, but the terms are not interchangeable. You can’t take information about Black voters and say it represents information about all voters of color. That’s just a convenient way for someone to erase groups of people like Latinxs. Not specifying that you’re referring to individual groups is also too often used as a way of co-opting other groups’ narratives without having to be accountable to them. That’s what’s happening here.

Biden has a notable lead among Black voters, though it reaches a plurality and not a majority of those polled. Biden does not have a notable lead among Latinx voters, and there haven’t been enough polls of us to know much for certain. Biden may have a lead among AAPI voters, but these polls are so infrequent that pollsters can’t assure much accuracy and can’t observe shifts in preference over time. Nobody polls Native voters because the media figures there’s not enough interest, instead of it being their job to teach why it should be important – so nobody really knows who different Native groups will support.

It is true to say that Biden leads among Black voters in opinion polling. It’s also important to listen to why. Yet that same information cannot then be extended to apply to other groups simply because you lack information about those groups. A plurality among Black voters should not be represented as a majority of Black people, nor as a majority of all people of color.

Biden doesn’t hold the same advantage among Latinx voters that he holds among Black voters. It is not accurate to say he leads among people of color when what you mean is that he leads among Black people. Biden may win other groups’ support, and he may not.

What should make you wary of this whole line of thinking is that white allies often like to come up with narratives that describe what people of color will do, why they’ll do it, and what it will mean…without people of color even having done it yet. It allows a co-optation of our actions before we take them, usually to claim a support that we haven’t yet decided.

Look at two white candidates whose campaigns are battling over having a mandate from people of color. Biden is using Black support to claim he leads among people of color. Sanders is using Latinx support to claim he leads among people of color. Meanwhile, both of them only poll in the 20s among all voters of color. No mandate has been given. No support is clear. Nobody’s even voted yet.

There have been polls where either Biden or Warren lead among white voters at about the same numbers. Where are all the claims that either has white support wrapped up, or a white mandate? Where are the claims that white votes have decided already? They don’t exist because there’s no political advantage to claiming that one or the other has white support secure.

If Black voters favor Biden in polls yet are still pretty split, and Latinx voters are largely undecided, that makes us all pretty similar to white voters right now. Yet no one’s declaring that anyone has white voters sewn up. That would be perceived as ridiculous, despite white voters also having favorites and being more decided than Latinx voters are right now.

Yet when that same logic is applied to voters of color, alternative media pushes the narrative, mainstream media takes the bait, and suddenly too many white allies are acting like it’s a narrative to be taken seriously, instead of the exact same level of ridiculous.

Nobody has us sewn up. There are favorites in some communities, just like among white voters, but nobody’s applying the same narratives or making the same claims about them that they are about us.

We get it. We’re useful for your candidate’s narrative. But if you want more than that, if you actually want that support to develop once voting rolls around, your candidate has to be useful to our reality. Work less on the narratives you sell, and more on the policies that help us.

If you enjoy what you read on this site, consider subscribing to Gabriel Valdez’s Patreon. It helps with the time and resources to continue writing articles like this one.

The feature image comes from a Boston Globe opinion piece here.