Tuesday, January 2, 2024

My 2024 predictions. By Matthew Yglesias


www.slowboring.com
My 2024 predictions
Matthew Yglesias
11 - 13 minutes

I have been dreading the need to do my annual predictions post — so much so that I strongly considered not doing it this year — because I didn’t want to have to think about what I got wrong last year.

But when I reflect on what I originally wrote in “How to be less full of shit,” the point of a predictions post is that it’s not supposed to be fun. What’s fun is to say stuff that’s interesting and maybe a little insightful and edgy and then not ask too many questions about whether it was actually true. The point of making explicit predictions and then looking back on them is to temper that impulse a little bit and to interrogate your own real views and levels of confidence more clearly.

And the good news is I am getting better.

My first year’s predictions were terrible, but I got better in year two, and this year’s calibration is better yet. Still not great, but pretty good.

    Things I said I was 90 percent confident in actually happened 100 percent of the time.

    Things I said I was 80 percent confident would happen actually happened 92 percent of the time.

    Things I said I was 70 percent confident in happened 65 percent of the time.

    My 60 percent forecasts happened 72 percent of the time.

    My 20 percent forecasts came true 20 percent of the time. 

This is pretty good. The two main flaws are that I was under-confident about my confident predictions, and I was no good at distinguishing between degrees of moderate confidence.

The importance of calibration and disciplined forecasting is particularly on my mind right now, because I’m, of course, thinking about the 2024 election. I think it’s important to recall that during the 2020 campaign, Donald Trump very confidently predicted that if Joe Biden won, the American oil industry would be shut down, the murder rate would surge, and we’d have rioters burning down our cities. None of those things happened — in fact, Trump was directionally incorrect in all of those predictions and we had more oil, less murder, and fewer riots under Joe Biden. Of course in politics, lots of people make claims about their enemies that turn out to be overstated; certainly Hillary Clinton made some overstated forecasts of doom related to a Trump victory. But it’s striking when people make forecasts that actually go in the wrong direction. It suggests there’s something fundamentally inaccurate in the conservative movement’s understanding of the underlying forces working on the policy areas that they themselves say are important.

So before we get to this year’s predictions, let’s look back and see what we can learn.

Here’s a fun example of the difference between good epistemics and good content. I wanted to put some sports predictions into last year’s list, so I just googled around on the internet for 10 different sports outcomes where the betting odds were about 20 percent. Some of those were related to a sport I actually follow closely (the NBA) or at least am broadly familiar with (the NFL), but two of them were related to the UEFA Champions League. And the only reason I even know what that is that after registering the official predictions, I looked it up.

The upshot is that I was perfectly calibrated, with exactly two of the things I said had a 20% chance of happening coming to pass:

    False: Celtics win finals (20%)

    False: Grizzlies win finals (20%)

    False: Bucks make finals (20%)

    False: Bruins win Stanley Cup (20%)

    False: Avalanche win Stanley Cup (20%)

    False: Bills win Super Bowl (20%)

    False: Eagles win Super Bowl (20%)

    True: Chiefs win Super Bowl (20%)

    False: Bayern Munich win Champions League (20%)

    True: Manchester City win Champions League (20%)

Of course, you don’t get a result that on the nose without a bit of good luck. But it illustrates that trusting the conventional wisdom can be underrated — especially in circumstances where people have actual money on the line in forming the conventional wisdom. I had no particular reason to believe that Manchester City had a one-in-five chance of winning the Champions League, but if sports gamblers say it’s true, then that’s pretty reliable.

None of this changes the basic reality that I have no value as a sports pundit, but I at least did an okay job of strategically manipulating my stats.

I made three different predictions about inflation, one of which was a bad miss:

    True: Inflation below 6 (80%)

    True: Inflation below 5 (60%)

    False: Inflation above 4 (80%)

In retrospect, I was under-confident about the Fed’s determination to bring inflation down, especially because I also wrongly predicted that the stock market would be down on the year. I should have thought that through better. My thinking on the stock forecast was that the Fed would be willing to really slam growth expectations in the name of bringing inflation down. But if that’s what I thought (and I think it was the right thing to think), then I should have been more confident on the Fed driving inflation down below five and more open to the possibility that it would get below four.

My other economics and politics predictions were pretty good, albeit under-confident, with one glaring exception: I said that Ron DeSantis would be leading the polling for the GOP nomination. I’m honestly not sure why I thought that, the guy is too short to be president.

Here was a more subtle miss:

    True: At least two bills pass that violate the “Hastert Rule” (80%)

    False: Kevin McCarthy ends the year as Speaker of the House (70%)

I don’t think it was crazy to think McCarthy had a decent chance of surviving conditional on violating the Hastert rule, but I was under-confident about the need to break the Hastert rule, which led me to be too bullish on McCarthy’s chances of making it through the year.

I made a bunch of international election forecasts and did a decent job of predicting that a lot of incumbents would lose. But I displayed some left bias and predicted an incumbent win in Berlin and an incumbent loss in Alberta, neither of which came to pass.

This group of predictions was more interesting:

    True: No Chinese attack on Taiwan (90%)

    True: No new countries join the European Union (90%)

    True: No countries leave the European Union (90%)

    True: No nuclear weapon use (90%)

    True: No peace deal in the Russia-Ukraine war (80%)

    False: Turkey and Hungary approve the accession of Sweden and Finland to NATO (70%)

    False: This year Israel and Saudi Arabia really do normalize relations! (60%)

    True: Biden visits Vietnam (60%)

    True: Some further easing of sanctions on Venezuela (60%)

    False: The U.S. and Canada really do reach a deal on the softwood lumber dispute (60%)

Obviously the 10/7 attacks and Israel’s response (in part, likely, by design) helped scuttle that Saudi-Israel diplomacy, which seemed like it was going pretty well previously. But perhaps another bias of mine is that when I see win-win deals on the table, I’m inclined to expect the deal to get done. It’s really annoying that neither NATO accession nor the softwood dispute has been worked out.

All things considered, though, I feel pretty good about my predicting.

Last year, I had a bunch of correlated blown calls that involved underestimating the political impact of Dobbs. This year, I think I did a better job of seeing the whole field. I don’t know about 2024, though, because I think it’s hard to stay objective about such a big American presidential election that’s going to correlate with a lot of other political outcomes.

Y’all already know my secret here, so I’m just gonna write these down and not say much about it:

    Denver Nuggets win NBA championship (20%)

    Philadelphia 76ers make the NBA finals (20%)

    Phoenix Suns make the NBA finals (20%)

    Luka Doncic wins the NBA MVP (20%)

    Baltimore Ravens win the Super Bowl (20%)

    Philadelphia Eagles make the Super Bowl (20%)

    New York Rangers make the Stanley Cup finals (20%)

    Las Vegas Golden Knights make the Stanley Cup finals (20%)

    Bayern Munich wins the Champions League (20%)

    Real Madrid wins the Champions League (20%)

Go Bayern!

We’re clearly on track for a good election for Republicans next year.

Whenever I point that out to friends, they push back, saying that a lot can change, that polls a year out aren’t predictive, that economic perceptions are likely to improve, and other things that are certainly true. What I think people often forget, though, is that “things can change” is symmetrical. Biden’s tenure as president has been marked by a number of somewhat unexpected pieces of bad luck — starting with the way the SARS-Cov-2 virus turned out to mutate rapidly so that vaccination didn't create an unambiguous end to the pandemic.

I’m not especially predicting any unexpected bad news for Biden (or America) next year, just trying to remind people that while the president could absolutely catch some lucky breaks, he may also catch some unlucky ones.

    Donald Trump is elected president (60%)

    Republicans maintain their majority in the House (60%)

    Republicans gain a majority in the senate (90%)

    Democrats lose at least one senate seat outside of WV/MT/OH (60%)

    Kelly Ayotte elected governor of New Hampshire (60%)

    Josh Stein elected governor of North Carolina (60%)

    No Biden change in Biden’s cabinet (80%)

    Donald Trump is the Republican nominee (90%)

    Joe Biden is the Democratic nominee (90%)

    No Supreme Court vacancy (90%)

    Immigration/border/asylum legislation passes (70%)

    At least one government shutdown (60%)

    Pro-GOP electoral college bias diminishes from its 2024 level (80%)

    Net electoral college bias continues to favor the GOP (60%)

    No split between the popular vote and the electoral college (80%)

Economically, though, I do think things will be looking up:

    Fed cuts interest rates at least fifty basis points (90%)

    Fed cuts interest rates at least 75 basis points (70%)

    Fed does not cut interest rates by more than 100 basis points (70%)

    Last CPI print of the year shows annual inflation below 3% (80%)

    Last CPI print of the year shows annual inflation above 2% (80%)

    No recession in 2024 (90%)

    Positive RGDP growth for each of the next four quarters (Q4 2023 - Q3 -2024) (70%)

    Unemployment rate stays below 4% (80%)

    Unemployment rate stays above 3% (80%)

    S&P 500 is up for the year (80%)

    US oil production hits a new record in 2024 (90%)

    GPT-5 is released (90%)

    Prime age EPOP falls short of its year 2000 peak (70%)

    New housing starts average over 1,500 per month (70%)

    Food prices more expensive in December 2024 than December 2023 (60%)

    African National Congress loses its majority in the South African National Assembly for the first time ever (60%)

    Claudia Sheinbaum elected President of Mexico (70%)

    Nicolás Maduro re-election (or “re-elected” if you prefer) in Venezuela (80%)

    KMT wins the presidential election in Taiwan (60%)

    Social Democratic Party (the center-right party) leads the next government in Portugal (70%)

    Vladimir Putin re-elected in Russia (90%)

    Far-right Freedom Party will be largest in Austrian Parliament (70%)

    Nayib Bukele re-elected as president of El Salvador (90%)

    Belgium won’t be able to form a government between their election and the end of the year (70%)

    Alexander Stubb elected President of Finland (70%)

Feel free to share your own predictions in the comments, and check back next year to see if I’m right about anything!

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.