Turnout was a key part of Democrats’ wins in 2018 – and is likely to be important in 2020 and beyond. Even a small bump in a reliably blue-voting group, in a key state, could make the difference in 2020. And longer term, the population is getting bluer, so if we maximize registration and turnout, the country will ultimately move to the left.
So, it’s reasonable to argue that we should focus a sizable amount of effort (and dollars) on boosting turnout in 2020 – but how? Which tactics (canvassing, direct mail, email, etc.) give the most “bang for buck”? What’s the maximum impact on turnout we can reasonably expect from these efforts? And should we focus on registration, GOTV, or both?
Our turnout/GOTV analysis in this post is part of our longer-term look at Dem strategies in 2020 and beyond. In the coming weeks, we’re going to dig into ways Democrats can win in high-priority states, by looking at the sizes and voting trends (“Dem votes per voter”) of different voter subsets (see here and here for our preliminary analyses), and assessing the extent to which turnout, persuasion, or both are key for success (see here for context).
This turnout-focused post summarizes current thinking and data from three main sources: a fabulous book, Get Out The Vote, by political scientists Donald Green and Alan Gerber; a great online resource, the U.S. Elections Project, maintained by political scientist Michael McDonald; and the U.S. Census Bureau. We take full responsibility for any misinterpretations or misrepresentations of these folks’ great work.
In this post:
What are current and historical voting rates in Presidential years?
What’s the maximum turnout we might hit in 2020?
Which GOTV tactics are most effective?
What about registration?
How can we make a difference in voting and registration rates in 2020?
1. What are current and historical voting rates in Presidential years?
The U.S. Elections Project (USEP) calculates voting rates based on the voting-eligible population (VEP), which takes into account individual states’ rules about parolees, etc. According to USEP, there appears to be an underlying trend of increased voting rates. (Full data and a nice graph here.) Turnout was 60.1% of VEP in 2016. That’s a bit down from 2008 (61.6%), but still on a trend line that’s been growing since 1972, when it was only 56.2%. (Importantly, these “voting rates” depend on not just turnout, but also registration – we’ll talk more about that later on.)
Our key takeaways:
Voting rates seem to be trending upward over time since 1972.
We think that means there’s likely an underlying “momentum” for higher voter turnout, independent of specific GOTV activities.
Successful GOTV efforts could accelerate voting rate growth even more – although the data here are scant, and it’s hard to know exactly how much of a boost to expect.
2. What’s the maximum turnout we might hit in 2020?
Based on historical data, we think it’s reasonable to expect a several-point bump in turnout in 2020 in subgroups of voters that are critical for Democrats, if we’re smart about deploying proven GOTV strategies.
The voting rate trend line since 1972 is clearly increasing over time, so we think there’s every reason to expect that the wind is at our backs, and that turnout should rise in 2020 compared with last election.
How high could it go? Looking at 2008 and 1968, total voting rates in both years were higher than they were in 2016 (by 1.5% and 2.2%, respectively) – so we know that those levels are feasible. It seems plausible to us that we could hit comparable levels in 2020 with a “base case” amount of GOTV activity.
Could we drive even greater turnout in particular subsets of voters? We think so, based on past data. For example, the voting rate among non-white Hispanics was 46.5% in 2008, but only 44.9% in 2016. Similarly, turnout in the 18-29 age group was five points lower in 2016 than in 2008 (43.4% vs. 48.4%). Again, these higher turnout numbers are clearly achievable, so it seems plausible to think we could hit them again, and maybe even exceed them. (And let’s be clear – even returning to 2008 levels would be a big boost for Democrats in 2020!)
Our key takeaways:
Hitting 2008 turnout levels in 2020 among key groups would make a pretty big difference for Democrats.
This seems plausible / feasible with a reasonable “base case” level of activity to support GOTV and mobilization activities.
As an “upside,” we think smart, targeted GOTV investments could capitalize on the forward momentum that seems to exist in turnout, which may let us hit total even higher voting rates in key subgroups.
3. Which GOTV tactics are most effective?
The data from turnout experiments (summarized in Get Out The Vote) are pretty conclusive – the most effective GOTV tactics per dollar are from “personal touch” techniques, especially face-to-face canvassing and phone calling by highly motivated volunteers (but not by commercial firms). Further down the list are things like mailings and texts.
Here are a few things the data tell us about canvassing:
Canvassing works – In 49 of 56 published experiments, canvassing increased voter turnout. There’s more high-quality data supporting canvassing than pretty much any other GOTV technique.
Canvassing is cost-effective – A basic model suggests a yield of one additional vote per 16 contacts, not including spillover effects on housemates. Even if you have to pay canvassers $16 per hour, at 6 contacts per hour you’ll get one additional voter per $33.
Volunteer phone calling is also pretty good, but the data are less uniform – On average, callers can get one additional voter to show up per 36 contacts. That’s pretty much equal to canvassing if you assume higher efficiency (16 contacts per hour). The wrinkle is that the yield is highly dependent on the motivation level of the callers; at the low end (i.e., using commercial firms), it’s only one voter per 106 contacts. For context, the average yield from texting is one vote per 312 targeted voters.
Canvassing is particularly well suited for Dems in 2020 – In a high-awareness election, canvassing has its highest yield among low-propensity voters. (The reverse is true in low-awareness races.) That means that there’s a lot to gain in 2020 from targeting groups with lower voting rates, like non-whites and youth, with face-to-face GOTV efforts. We suspect the same is true for volunteer-led phone calling too.
Our key takeaways:
Face-to-face canvassing and personal phone calling are key strategies for voter mobilization for Dems in 2020.
In a Presidential year like 2020, “personal touch” approaches are particularly well suited for boosting voting among lower-turnout groups.
These high-touch methods are extremely cost-effective, but they need a lot of person power to maximize their impact, as well as a strong organization to provide training and vet the target lists.
4. What about registration?
Many (but not all) studies show that GOTV has its greatest impact in the week before Election Day. (That doesn’t count early voting, obviously.) Early donations help candidates and non-profits build the organization and data they’ll deploy in the final pre-election surge – so giving to them in these early days is important.
But besides readying ourselves for GOTV efforts, another thing we can do in the run-up period is maximize voter registration. In states with pre-registration requirements, the deadline is about a month before the election. That gives us plenty of time to make sure that folks who want to show up to the polls are legally able to do so.
How much does registration matter? It’s a hard question to answer. Just counting the number of registered voters is problematic, because voter rolls contain many folks who are “inactive” (due to moving, for example). The Census collects registration data from a survey, and although this is imperfect information, it’s better than nothing.
From Census survey data, about 14% of voting age citizens report not being registered. (Another 15% refused to answer the question, weren’t asked during the survey, or said they didn’t know.) But for non-white Hispanics, 24% explicitly said they weren’t registered, and another 19% didn’t respond. Similarly, focusing on the 18-24 age group, 25% said they weren’t registered, and another 20% didn’t respond. So whatever the “real” number of eligible but unregistered voters is, it appears higher in several groups, like non-whites and youths, that tend to side with Democrats.
We also have some information about registration tactics from experimental evidence. As is true with turnout, it seems like personal appeals, like in classrooms and neighborhoods, work pretty well to register voters compared with less personalized methods (except for mailings from state agencies). We also know that sending registration materials to folks turning 18 works pretty well to boost registration. And finally, data show that especially in an election year, a sizable fraction of new registrations turns into new voters.
Finally, let’s also be clear that we need to keep focusing in parallel on policy changes that will boost registration, like “motor voter” laws. As we get closer to November, we’ll highlight key state elections (for Secretary of State and legislators) and ballot measures that we need to support to cement long-term gains in enfranchisement.
Our key takeaways:
It’s hard to precisely quantify the likely impact of registration efforts on voting numbers for 2020.
That said, there are clearly larger registration gaps among several key constituencies for Democrats than for the population at-large.
Many of the same tactics that work well for GOTV also help boost registration, although the “bang for buck” is harder to calculate.
5. What can we do?
We think all of this leads to three tactical things we can do to help boost turnout for Dems in November:
Support local candidates in key regions: State legislature and House campaigns tend to spend a lot of effort and resources on volunteer-led local canvassing and phone calling – which are the most effective GOTV and registration tactics! The turnout that these campaigns drive will almost certainly “trickle up” into blue votes for Senate and President. Plus, state legislature races are critical in their own right to combat gerrymandering and other forms of voter suppression. (See our prior post on that topic here.)
TAKE ACTION: Tell your friends to subscribe to our blog (and subscribe yourself if you don’t already)! In the coming months, we will cover crucial House and state legislature campaigns in key states, and give you clear advice on where we think we can make a difference in 2020.
2. Support GOTV and registration organizations that focus on in-person tactics: On a broader level, organizations that are targeting Dem-leaning subgroups in key areas can have a big impact in 2020, especially if they’re focused on canvassing and personal calls.
TAKE ACTION: At the national level, we are fans of When We All Vote, HeadCount, Voto Latino, and local chapters of the League of Women Voters. (If you want to suggest more GOTV/registration groups we should list here, please email us!)
3. Volunteer for phone banks: We were pleasantly surprised to learn that volunteer phone banking works almost as well as in-person canvassing – great news for folks who live in deep-blue states, but want to make a difference! Let’s put this in perspective: in 2016, Clinton lost the Wisconsin popular vote by about 23,000 votes. We previously calculated that for each new young college-educated woman in Wisconsin who shows up to the polls, we’d expect a net gain for Dems of 0.38 votes. (See our earlier post for details.) In other words, we could move the needle by getting about 60,500 more of these folks to show up on Election Day. At 16 calls per hour, this equals under 4,000 volunteer hours of phone banking. We can do this!
TAKE ACTION: Besides reaching out to specific campaigns, check out the great work that Sister District is doing to help motivated telephoners make a difference!
Want to read more from Blue Ripple? Visit our website, sign up for email updates, and follow us on Twitter and Facebook. Folks interested in our data and modeling efforts should also check out our Github page.
(Image: MIT News)