Q&A #1: Presidential primaries, polling and forecasting, and inflation and the midterms

Welcome to the blog’s new monthly(-ish) Q&A. Remember that you can submit questions for future editions here.

Readers sent in a lot of great questions to answer this week. I have done my best to select one big topic (on how parties ought to pick presidential candidates!) that I think all readers will enjoy, and picked a few other questions (which today just so happen to be a bit more technical) to put behind the paywall.

If you enjoy this post, please share it!

---

Doug asks: “After 2020's pervasive 15% threshold nearly preventing Democrats from rallying behind an electable presidential candidate, any serious thoughts about reform before 2028?

This is a good question, and interestingly one of the few backward-looking ones on the submission sheet.

Setting aside what “electability” means or whether we can even directly measure it, what I thought of first was how the 2020 Democratic Primary has a reputation for being atypically focused on electability. But I’m actually not sure that the party is meaningfully more electability-minded than before. In a Gallup poll from June 2019, 58% of Democrats said they preferred a candidate with the best chance of beating Trump over one who they agreed with on most of the issues they care about.

In comparison, the share of Republicans who prioritized electability over issue congruence was 53% for primary voters in 2011 and 55% for Democrats in 2004. Given Gallup’s large sample sizes those numbers are likely statistically significant, but they do not strike me as meaningfully larger than another — eg in terms of whether one candidate would have had a chance in 2004 or 2020, holding everything else about the party and election circumstances equal.

OK, so, Democrats were focused on electability, but maybe not any more than they or Republicans usually are. That makes sense given the degree to which partisans want to avoid victories by their political opponents. About 90-95% of Democrats and Republicans vote for the same party year after year. It is reasonable to think the vast majority of Democrats would have voted for Bernie Sanders or Elizabeth Warren — less “electable” candidates — and, similarly, that Republicans in 2016 would have also turned out at about the same numbers for a Marco Rubio campaign. (NB: One note is that close general elections are fought on those 5-10% margins!)

What’s interesting to me here is how America has ended up with a unique, not-repeated-anywhere-else-in-the-world presidential primary system that somewhat emphasizes the tension between these two qualities (electability and representation) among nominees.

The process (at least on the Democratic side) of democratizing and devolving candidate selection to a proportional popular vote at the state level creates an array of weird incentives acting on voters and candidates. These incentives influence our decisions and, thus, the outcomes we get from the system.

One example comes from the spacing out of primaries across a primary calendar. This causes an overweighting of candidates who perform well among voters in early contests, who tend to be whiter, more educated, and more engaged than Democrats who vote later. And while candidates can overcome this — Joe Biden is arguably president because of good old fashions politicking among African Americans — we can imagine worlds where outcomes are substantially different. So, one idea for reform is to compress the calendar — or just have all the voters cast ballots on a single day.

This is going to get messy when you only have 20 candidates. So the second proposal is to allocate votes to candidates using ranked-choice-voting, rather than assigning convention delegates to them proportionally (and only if a candidate receives more than 15% of the vote). A nationwide primary using ranked-choice voting would have the benefit of balancing better between ideological representation and electability. So long as you think people should be picking party nominees, that would be an improvement over the current system.

There is one outstanding criticism, however. Political scientists and presidential histories have been particularly focused recently on what the utility of democratized candidate selection is, anyway. One theory is that the system makes voters happier with their nominees; with a party primary where delegates are bound to candidates, you never get a Democratic nominee Hubert Humphrey (who did not enter any state primaries before his contentious nomination in 1968).

But democratization may also open the door to polarization and extremism, especially on issues of democracy and electoral liberalism. I am drawing here on an argument made by Elaine Kamark in her book Primary Politics. In democracies overseas, for example, parliamentary leaders are selected by representatives, not voters; then, voters get to cast indirect up-or-down votes for them via the general election process. Kamark proposes a variety of potential reforms, including letting elected representatives nominate the candidates that people then vote on. They act as useful and informative gatekeepers, in her view.

Of course, if you’re not happy with the major positions of either the Democratic or Republican parties in America, you are not going to be happy with a system that even further removes people from the process. But then your complaint is really with elections that assign winners via plurality vote, not with the particularities of party primaries.

That makes three reforms:

1. Compress the primary calendar

2. Use ranked-choice-voting (perhaps with multiple winners, called “STV” in the reform community) so eventual tallies better represent the balanced concerns of voters

3. Give party leaders a heavier hand in deciding who gets to run on their ticket. (You may avoid Trump by doing this.)

---

OK, that was a long answer. I’ll try to be faster with these next ones:

Seth D asks: “Given the GOP overperformance at the ballot box relative to polls in 2016 and 2020, how much should we round up on the Republican side when we see polling about 2022? are you spotting them a couple of points mentally?”