May 1Liked by Robert de Neufville

Interpreting the keys is subjective. Based upon Lichtmans book, the third party key is flipped if a candidate gets 10% or more and RFK has. RFKs is also the primary that the Democrats tried to quash. He was trying to run in the Democratic primary. Also, Trump and Biden can both be categorized as incumbents. People remember that Trump was President before. 70% say the economy is bad. Afghanistan withdrawal and the current wars can be viewed as foreign policy failures. Lichtman is currently highly biased in how he interprets the keys. If you want to be impartial. Consider how it is possible to interpret a key and the reason for it as flipped. A technical recession or a majority of people experiencing a bad economy, more inflation and harder to make ends meet and afford the basics. A policy that passed or one that a majority really care about.

Expand full comment

Interpreting the keys is subjective—that's definitely one of the problems with using them as a predictive tool—but Lichtman's interpretation seems reasonable to me. I'm fairly confident RFK won't be polling at 10% closer to the election (and I'm actually skeptical that RFK will siphon significantly more votes from Biden than from Trump). You certainly can view the withdrawal from Afghanistan or the wars in Ukraine or the Middle East as policy failures, but I think it's clear they're not perceived by most Americans as disasters in Lichtman's sense. I do agree both that Trump is like an incumbent in some sense and that perceptions of the economy are real problem for Biden.

Expand full comment

I have gone through the 1984-2020 elections which is when the 13 keys existed. the 1980 and prior was overfitting to create the 12 and 13 things to explain the history up to that point. The reasons sound reasonable. https://www.metaculus.com/questions/11245/winner-of-2024-us-presidential-election/#comment-153510

1984 Mondale was a boring shadow of Carter and Carter had lost 4 years before. Blamed for high inflation. Lost by 16%. People could compare 1976-1980 vs 1980-1984. not 13 keys dependent and any method would work for the obvious pick.

1988. Dukakis was doing good until the Sept 18 tank photo. the polling shows he was sometimes leading and then tank photo and he was a laughing stock and lost 20%. Not a 13 key.

1992. Bush Sr was actually doing good, leading. then in June he flip flopped on his core promise and first term campaign slogan. Read my lips no new taxes. 20% drop. So the 13 keys would have treated him and Dukakis the same until those Earthquake events. 13 keys is based upon known earthquake events 1776-1980.

I go through all of the rest at the link.13 keys adds little to the analysis. There needs to be more negative weight on being boring, unlikable and a poor campaigner. Kerry, Romney, McCain, Hillary. If we take the blowouts off the table, then most of the last elections from 2000 are decided at key states and electoral college.

80% support for Israel but Democrats lose by 3% in Michigan unless a hard turn against Israel was made but a hard turn does not result in winning Michigan, just losing with a different mix. There are no policies or campaign words that thread the needle to flip the needed majority of battleground states.


Meanwhile, erosion of support. Inflation already bad, keeps worsening. Core PCI trending up. But even if it was not people pay 20-30% more for food every day. 50%-double at restaurants. Even rich people I know get sticker shock at restaurants and complain about it. Everyone goes out less. 7% mortgages mean homes are not affordable but not just that even the hope of someday affording a home seems to be slipping away.

It is the large increase in even high income people living paycheck to paycheck or in debt.


18% lived paycheck to paycheck in 2019, now 36% live paycheck to paycheck.

What amount of spinning some perception would convince people? People were saving and investing some money in 2019 but now 20% more have gone negative every month. Savings have burned up. They have gone into debt at high rates. Tell them it is just their imagination, that they don't really have a problem.

The old 1980s, 1990s version of CPI would have recorded inflation peaking at 15% and not 9%. Change how we report the numbers but people still feel it.


This is why 55%+ say they are personally worse off under Biden.

15-20% had to take more loans, more credit card debt.

Hey unemployment is not that bad. How many are working 2-3 jobs to try to stop sinking into more debt?

Not even connecting on what is being actually experienced? Ignore the Trump part of it. Just look at the track record and what is happening to a large segment of the people.

Expand full comment
Apr 25Liked by Robert de Neufville

I would give biden 10/13 of the key's missing out on Charisma (x2) and major policy changes.

I tend to rely more heavily on a polling based approach, which currently puts me at Trump 55% Biden 45%. the 7 candidate tipping point states go as follows (negative means biden favor). Election polls this far out aren't super accurate but they aren't extremely inaccurate. I do agree that more "fundamentals" based approaches are more valuable method than it would be in october, but I still put >50% of the weight on the polls at this stage.

Pennsylvania -0.1

Wisconsin +2

Michigan +3

Georgia +4

North Carolina +4.8

Nevada +4.8

Arizona +5

Trump's electoral college lead is underrated IMO.

Trump's legal issues previously raised his popularity, why would you expect that to change in the future?

Expand full comment

Thanks for the comment. Here's Lichtman's analysis of where the keys stand this year:


I wouldn't say the polls are inaccurate right now. I think they measure voting intention fairly accurately. But I think most people's voting intention in April—before many of them are paying close attention to the election—is at best a weak predictor of how they will vote in November. I think it's fairly predictable (because he's the incumbent, because the Republican platform is unpopular, because of Trump's legal issues, because of Republican Party dysfunction) that barring some new event Biden will do better than he is currently polling.

I know some people believe Trump's legal problems have made him more popular, but I don't think that's really true. I think they energized his base, which helped wrap up the primaries quickly. But there's a lot of polling that suggests that if he's really convicted voters who might otherwise vote for him are likely to stay home. I also think—although I'm skeptical he'll actually be convicted of a felony in the so-called hush money trial—that it hurts his image as a dominant winner that he's being forced to sit in the dock right and defer to a judge right now.

I could certainly be wrong. We'll see in November.

Expand full comment
Apr 28Liked by Robert de Neufville

Great analysis. I agree with much of this.

However, if predicting now, I think Biden is in to squeek by again in a close election.

I think the keys are a good starting point, but in reality, each question isn't binary, and there are factors which lay outside of the keys. The economy is pretty great, but b/c of what happened with the Covid economy and inflation, Biden doesn't appear to get credit for it. I don't think polls at this stage are completely worthless, particularly all the polls that show that Biden's approval rating is consistently low relative to predecessors. Incumbents around the world post-Covid seem to be unpopular. The polls currently show Biden losing. Biden is running against another former President, and former Presidents tend to gain in popularity after office b/c they aren't attacked. As the campaign heats up, and Trump is attacked more, I do expect Biden's standing in the polls to be more likely to improve than otherwise. I'm long on a Biden victory. Another factor that lays outside the keys: Biden is in fact too old to be President again. This is probably going to make a landslide impossible, in addition to what looks like any of the consequential court cases to be pushed after November (there are so many I'm not sure about this). The one case that is now on is certainly not going to dislodge support among MAGA'ers. My view is that the case looks like weak sauce, and Trump may even win it, although predictors whom I respect thinks it is most likely he will lose. That said, a conviction may still sway some indies

Another thing -- Manifold, Metaculus, and the betting markets -- Insight Prediction, Polymarket, etc. all have this election very close. I do suspect they may overweight the polls. Trump even recently was a massive favorite over Biden.

Also, these two candidates have already run against each other once, which is surely the best base rate for this election, and that election was also very close -- not a 90% thing. Betting markets had that election closer than the polls, and in that one election at least, the average betting market did seemingly beat the models.

Expand full comment

You make some excellent points. I agree with you that this election doesn't seem like a 90% thing. I personally think Jean-Pierre is too confident Biden will win, but he is always among the most accurate Good Judgment superforecasters so I take his opinion seriously even where it differs from mine.

I do think it's reasonable for voters to be unhappy with the economy, although presidents normally don't deserve too much credit or blame for its performance. Prices are still higher than they were several years ago and mortgages are very expensive, so in spite of rising real wages some important things are less affordable. However, I think it's also true that a lot of the unhappiness with the economy comes from Republicans—who Biden won't win anyway—who would be unhappy with any economy under Biden (Democrats' feelings about the economy don't seem to be colored by who's president to the same degree).

I also think it's reasonable to be concerned about Biden's age. I think it is a consideration fro persuadable voters. But it's not obvious to me whether it is really a cause of Biden's unpopularity or largely an after-the-fact rationalization of it. I think one reason people focus on Biden's age is that it is harder to make other attacks (that he is corrupt, unamerican, extremist etc.) stick to a moderate, seemingly decent, white man. He gets attacked for his age in part because no one believes he is cryptomarxist who grow up in a Kenyan madrasa.

I'm not sure about the New York case. The case for him committing misdemeanor fraud seems strong, but the case for finding he committed a felony—which is what would really hurt him—seems weaker, but I'm certainly not an expert in the finer points of New York law. But win or lose I think it hurts him to have to sit in a courtroom and submit to a judge's authority. It hurts the authoritarian brand that is responsible for so much of his popularity.

I don't take the betting markets very seriously as a signal. I think they are very sensitive to the polls and do the general vibes of reporting. I also think that negative partisanship means that incumbent presidents are all likely to be unpopular under normal circumstances. But, that said, I agree it's significant that Biden has been as unpopular as he has been. If the polls go much longer showing Trump in the lead in the battleground states, I'll definitely have to reassess my forecast.

Expand full comment