clock menu more-arrow no yes mobile

Filed under:

The 14 predictions that came true in 2023 — and the 7 that didn’t

The 21 forecasts we made in 2023, revisited.

An illustration filled with bright yellow check marks and dark blue X marks. Also included throughout the scene: SpaceX Starship, Joe Biden, the top of Donald Trump’s head, lab-grown meat, the flag of Finland, the flag of NATO, the U.S. supreme court building, and a syringe with antibiotics. Paige Vickers/Vox

Though the name might suggest otherwise, we are not technically in the “predicting the future” game at Future Perfect. We usually leave that to the pundits and analysts who will confidently tell you about who the next Republican nominee for president or NBA champion will be — and then conveniently forget should those predictions fail to come true.

But there is real epistemic value in not just trying to predict what’s to come, but putting a specific probability on that prediction — and then, just as importantly, evaluating whether and why you were right (or wrong) after the fact. It’s an intellectual exercise in both rigor and humility, and one that is becoming increasingly valuable in our part of the media.

So how did our 2023 forecasts do? Not bad — 14 correct predictions to 7 misfires. (Note that we had to invalidate two predictions from 2023’s list, on the number of poultry culled because of bird flu and Beyond Meat’s stock price, because of problems in how the predictions were formulated.) Politics proved relatively easy — yes, Joe Biden would run for reelection and would remain the Democratic frontrunner; no, not a single Republican would seriously challenge Donald Trump’s hold on the party and the likely nomination.

Economics proved more difficult, as we and just about every other analyst failed to foresee that the US would escape recession even as it brought down inflation. And tech turned out to confound some of our expectations, in part because technologies like lab-grown meat haven’t advanced as rapidly as we’d forecast, and in part because bad things, like the ongoing avian flu outbreaks, haven’t been quite as bad as we thought.

As I do every year, I’ll quote my colleague Dylan Matthews: “Predicting the future is a skill at which some people are dramatically better than others, and practicing is one of the best ways to improve at it.” Check back with us on January 1, when we unveil our predictions for 2024. —Bryan Walsh

United States

Joe Biden will be the frontrunner for the Democratic nomination heading into 2024 (70 percent) — RIGHT

Naturally, 2023 featured a lot of speculation and suggestions about dramatic change-ups on the Democratic side: should Biden even run again? Should he replace Kamala Harris with Michigan Gov. Gretchen Whitmer? Would a top-tier alternative like California Gov. Gavin Newsom challenge him?

Ultimately none of that happened, and the strongest challenger he got was Rep. Dean Phillips of Minnesota, which must be who pollsters mean when they ask respondents their views of a “generic Democrat.” As a result, the prediction markets and platforms, from the crypto-based Polymarket to the staid and professional Metaculus to the goofy and anarchic Manifold to the old standby PredictIt, give Biden overwhelming odds to be renominated. This was the standard I chose for determining if Biden was the “frontrunner,” so I feel like I nailed this one. And, for what it’s worth, the polls agree. —Dylan Matthews

Donald Trump will be the frontrunner for the Republican nomination heading into 2024 (60 percent) —RIGHT

While a year later it feels inevitable that Trump would be crushing his rivals and that the Ron DeSantis bubble would’ve popped almost immediately, this was not exactly obvious in late 2022, which accounts for my relatively unconfident prediction. Trump, after all, was under investigation by several prosecutors, seemed likely to be indicted by a few of them, and is (no less than Joe Biden) showing his age these days.

But you can never go broke betting against Trump in a GOP primary, even after he was indicted several times, and so Polymarket and all the rest still put him as a decisive frontrunner as of this writing. —DM

The Supreme Court will rule that affirmative action is unconstitutional (70 percent) — WRONG

This may seem like a case where I was obviously right — the Court did rule affirmative action unconstitutional in most cases — but you have to look at the fine print. Here’s how I characterized my prediction:

The reason I’m not more confident is due to a nuance [my Vox colleague Ian] Millhiser noted, which is that [Chief Justice John] Roberts appeared open to racial preferences at military academies, noting the federal government’s argument that the military needs a diverse officer corps to succeed. If such a carve-out is included in the ultimate ruling, my prediction here will be wrong: I’m predicting they’ll strike down affirmative action across the board at public or publicly funded institutions.

The Supreme Court did include such a carveout allowing for racial preferences at service academies. Here is Roberts, in footnote 4 of his opinion:

The United States as amicus curiae contends that race-based admissions programs further compelling interests at our Nation’s military academies. No military academy is a party to these cases, however, and none of the courts below addressed the propriety of race-based admissions systems in that context. This opinion also does not address the issue, in light of the potentially distinct interests that military academies may present.

In her dissent, Sonia Sotomayor interprets this as meaning “the Court exempts military academies from its ruling.” I predicted they would not do this, so I got this wrong. I apologize to the good people at Manifold Markets whom I confused on this. —DM

The US will not meet its target for refugee admissions this fiscal year (80 percent) — RIGHT

Unfortunately, this prediction was right. President Biden had set the refugee admissions target at 125,000 for fiscal year 2023 but ended up resettling roughly 60,000. Even getting halfway to the target proved just out of reach.

America’s resettlement infrastructure still hasn’t fully recovered from the Trump administration, which gutted it. Biden promised to restaff the government agencies that do resettlement and reopen the offices that had been shuttered, but advocates say the rebuild has been too slow.

Yes, the US has welcomed some groups — like Afghans, Ukrainians, and Venezuelans — but note that those who came to the US via the legal process known as humanitarian parole only get stays of two years. They don’t count toward the number of refugees resettled, as refugees are given a path to permanent residency. —Sigal Samuel

The US will slip into recession during 2023 (70 percent) — WRONG

I was very wrong on this one. (Bad for me, good for the commonwealth.) At this point, it now appears that the US economy will likely have grown by more than 2 percent over the course of 2023, which, clearly, does not qualify as a recession. Despite concerns that the Federal Reserve’s campaign to quash inflation through interest rate hikes would inevitably squash growth, the US economy remained startlingly resilient in 2023, outperforming expectations across the board. Fed Chair Jerome Powell couldn’t have set up the country for a softer landing with a warehouse full of Tempur-Pedic mattresses.

But if I was wrong, I wasn’t alone. Recession expectations were historically aligned — everyone from Wall Street analysts to Fed economists to intense guys who really want you to buy gold largely assumed a recession was an inevitability sometime in 2023. Heck, according to one survey, 59 percent of Americans feel like the US is in a recession right now, which is a whole other thing. (See above: it is not.)

You can’t really blame the prognosticators. The US has almost never managed to curb inflation at this level without slipping into a recession. Economically speaking, what has happened in 2023 is akin to water suddenly flowing uphill — which is probably why a lot of analysts are still worried about the possibility of a recession next year. We’ll see if Powell can pull another rabbit out of his hat. —BW

Inflation in the US will exceed 3 percent (60 percent) — RIGHT

My definition of inflation for my predictions is the same as the one used by the Federal Reserve: the price increases of “personal consumption expenditures,” excluding food and energy. More specifically, I committed to using an average of the first three quarters of the year, as the fourth quarter data is not yet available.

Well, the first three quarters’ inflation rates were 5.0, 3.7, and 2.3 (see row 34 here on page 12), for an average of 3.67 percent. That is, for sure, above 3 percent, even as it was rapidly falling. Even adding in October’s data results in average monthly inflation of 0.28 percent for the year, or 3.36 percent in annual terms.

I think 2023 will be the year that inflation finally gets back to the 2 percent range, which is an impressive achievement for the Fed given no recession has occurred. But for 2023, it was still fairly high. —DM

There will be no Supreme Court vacancies in 2023 (90 percent) — RIGHT

Given the nine justices’ ages, wealth, and education levels, there was a less than 11 percent chance that a sitting Supreme Court justice would’ve died this year. That didn’t happen, nor did any sitting justice retire.

That could change next year. There’s growing pressure for 69-year-old liberal Justice Sonia Sotomayor to leave the court before Republicans likely gain control of the Senate in 2024 and hold power over Supreme Court nominations for the next six to eight years, or longer.

That pressure will grow in 2024, in what’s already gearing up to be a wild ride of an election year. If Sotomayor were to die while Republicans control the Senate, it could lock in a conservative 7-2 majority for years and further erode American democracy. —Kenny Torrella

The world

Vladimir Putin will still be president of Russia (80 percent) — RIGHT

Putin recently announced his reelection bid for a fifth term as Russia’s president, and given the average life expectancy of Russian opponents of his regime, I give him very good odds.

My estimate of 20 percent odds of Putin losing power was based on a suspicion that the stalemate in Ukraine, and the massive economic and human toll it’s wreaked on Russia, would make him vulnerable. That was correct, and in June the mercenary Wagner Group and its colorful leader Yevgeny Prigozhin openly mutinied against Putin and began to march on Moscow. For a brief moment, it appeared they would be able to take the city and perhaps overthrow Putin.

But Prigozhin — who was not an opponent of the war but instead a believer that he could run it better than Putin — blinked and called off the march. Putin, at first, seemed to welcome him back into the fold. Then, on August 23, two months after the mutiny, the plane Prigozhin was flying in crashed due to an explosion on board. Putin has suggested that Prigozhin died when a cocaine-fueled hand grenade-tossing party aboard the plane got out of hand. As plausible as that seems, I agree with other analysts that it seems more likely Putin just killed the guy.

In any case, Putin was not able to prevent the pressures of the war from building into a dangerous mutiny. He was able to crush that mutiny, though, and to send a message that any future attempts will end in fiery death. —DM

China will not launch a full-scale invasion of Taiwan (90 percent) — RIGHT

It would be an exaggeration to say that Chinese relations with Taiwan are currently good. The Democratic Progressive Party — the more pro-independence, anti-Beijing party on the island — is currently leading polls for next month’s presidential election, albeit narrowly. China keeps sending carriers through the Taiwan Strait, and is reportedly meddling in the election to try to help the pro-unification Kuomintang party.

But there have not been any indications that China is amassing the troops it needs for a full-scale amphibious assault, or the ships it would need for a blockade meant to force Taiwanese capitulation. And thank goodness; the world hardly needs another high-casualty war right now. —DM

At least one new country will join NATO (90 percent) — RIGHT

Finland, which despite being a liberal democracy remained so diplomatically close to the Soviet Union that the country’s name became a term of diplomatic art, joined the North Atlantic Treaty Organization on April 4, in the most dramatic reaction of any country to Russia’s assault on Ukraine. The country, which was a Russian possession until 1917, shares an 832-mile border with Russia, which can now host NATO troops and bases from allied nations to deter Russian incursions west.

Sweden seemed likely to formally join this year as well, but it was blocked due to foot-dragging by Hungary and Turkey over Sweden’s criticism of Hungarian autocrat Viktor Orbán and its past refusal to extradite Kurdish activists to Turkey, respectively. It still seems likely to join in 2024, but its path is a little more circuitous than Finland’s. —DM

Finland will remain the world’s happiest country, while America won’t crack the top dozen (75 percent) — RIGHT

Every year, the World Happiness Report ranks countries in terms of the happiness of their populations. It’s part of a burgeoning movement to pay more attention to indicators of subjective well-being as opposed to just raw GDP.

This year’s country rankings didn’t surprise me at all. Finland held on to the top spot on the list, thanks to its well-run public services, high levels of trust in authority, and low levels of crime and inequality, among other things. I was pretty confident that would be the case because the Nordic nation had already been the happiest country for five years running, and last year researchers noted that its score was “significantly ahead” of every other country.

Meanwhile, America’s ranking improved very slightly — from 16th place in 2022 to 15th place in 2023 — but, as I predicted, it didn’t make it into the top dozen spots. It never has, which is, um, really something to reflect on. —SS

Science and technology

A psychedelic-based mental health treatment will win US regulatory approval (60 percent) — WRONG

Based on indications from experts and the government, I suspected there was a decent chance regulators would approve MDMA for treatment of PTSD this year. And after publishing some promising study results, the Multidisciplinary Association for Psychedelic Studies did indeed file for Food and Drug Administration approval of MDMA-assisted therapy for PTSD. But it’ll be months before we get a decision out of the FDA. —SS

The US will not approve a nasal vaccine for Covid-19 (90 percent) — RIGHT

Even though we’ve been told that Covid-19 vaccines delivered through the nose may prevent more infections than shots in arms, and even though Covid-19 nasal vaccines created by American researchers have been tested in animals, the US doesn’t have them and probably won’t anytime soon. One big reason is a lack of funding: Biden asked Congress for more money for next-generation vaccines, but Republicans resisted. —SS

An AI company will knowingly release a text-to-image or text-to-video model that exhibits bias (90 percent) — RIGHT

When I made this prediction, I specified that I would judge an AI company to have “knowingly” released a biased model if the company acknowledges in a system card or similar that the product exhibits bias.

Well, in October, OpenAI released DALL-E 3, and stated in the model’s system card: “Bias remains an issue with generative models including DALL·E 3, both with and without mitigations. DALL·E 3 has the potential to reinforce stereotypes … We additionally see a tendency toward taking a Western point-of-view more generally.”

Other AI models, like Google’s recently unveiled Gemini, almost certainly exhibit bias, too — it’s just that, unlike OpenAI, Google is not saying what’s under the hood. —SS

OpenAI will release GPT-4 (60 percent) — RIGHT

GPT-4, released on March 14, was not a dramatic sea change in ability compared to GPT-3.5 that preceded it. But it’s quite a bit better, especially combined with other improvements that OpenAI rolled out this year: Code Interpreter, which can generate working code to solve problems based on plain English prompts; DALL-E 3, the latest OpenAI image generation model now integrated into ChatGPT; GPT-4 Turbo, yet another refinement of the core model; and GPTs, a program that enables users to train their own custom version of GPT-4 tailored to a particular task.

The boardroom chaos that consumed the company in November seems, in retrospect, to be mostly a blip in the context of its big product releases. It remains by far the dominant AI company, and with the aggressively commercializing Sam Altman now more firmly in charge than ever, it shows no signs of slowing down. —DM

SpaceX’s Starship will reach orbit (70 percent) — WRONG

SpaceX certainly tried to reach orbit in 2023. It launched the Starship twice, on April 20 and November 18. The first saw the vehicle explode after reaching 39 km, and the second saw the second stage reach 148 km before a safety procedure led it to self-destruct. But neither entered orbit; even a successful launch, by SpaceX’s own standards, would not have led to a full orbit of the Earth.

I think my failure here was partly raising a poorly framed question. What I meant, I think, was “will SpaceX have a Starship test that goes well.” I think the November test went well by many metrics. But as I phrased the question, I set the bar implausibly high, and SpaceX failed to meet it. —DM

Animal welfare

At least three lab-grown meat companies will begin selling their products in the US (50 percent) — WRONG

Two lab-grown, or “cell-cultivated” meat, companies began selling their products in the US in 2023. I was wrong here, but I had hedged my bet with 50 percent confidence because I had heard so much uncertainty from people in the sector about which companies would first get approval from US regulators to sell their products, and when.

Two of the startups with the most funding, Upside Foods and GOOD Meat — both based in the San Francisco Bay Area — gained approval the same day in June. Both make chicken derived from chicken cells, which they feed a mix of sugars, amino acids, salts, vitamins, minerals, and other ingredients for several weeks until they can be harvested as animal fat and muscle tissue.

The startups are selling their products to consumers, but in very limited quantities at just one US restaurant per company. They still have a long way to go to figuring out if they can scale their technology to compete with conventional meat on cost. But overcoming the regulatory hurdle is part of the battle in bringing a product to market, and this nascent field demonstrated their processes are safe and regulatory-compliant. —KT

The Supreme Court will rule in favor of the pork industry in National Pork Producers Council v. Ross (70 percent) — WRONG

In 2018, California voters passed a law, known as Proposition 12, that requires pork sold in the state to come from pigs given more space — essentially, cage-free conditions — whether those pigs were raised in California or not. A pork industry group, the National Pork Producers Council, sued the state over it, and the case made its way up to the US Supreme Court. I predicted with high confidence that the business-friendly Court would rule in favor of the pork producers, but instead — to my shock and delight — they sided with California and the pigs.

The case hinged not on animal welfare, but on states’ rights, and whether Prop 12 was unfairly forcing farmers in other states to give pigs more space if they still wanted to sell their pork in California. All nine justices agreed Prop 12 was constitutionally sound in this regard. The pork industry also claimed that the financial trouble the law imposed on producers outweighed any benefits the law delivered to Californians. On this matter, the justices voted to uphold Prop 12 and scrambled the political divide on the Court. Conservative Justices Neil Gorsuch, Amy Coney Barrett, and Clarence Thomas joined liberal Justices Sonia Sotomayor and Elena Kagan, for the 5-4 decision.

It was a momentous decision for the future of farm animal welfare, helping to ensure similar laws around the country remain intact and giving animal advocates a level of certainty that it’ll be harder to challenge future laws. —KT

Antibiotics sales for farmed animals will increase in 2022 (65 percent) — RIGHT

Farmed animals are raised in unsanitary, overcrowded conditions. Knowing that disease spread is all but inevitable, meat producers routinely feed animals antibiotics. A terrifying result of routine antibiotic feeding is that bacteria are mutating and developing resistance to these antibiotics, making them less effective in treating common conditions in humans, like sepsis, urinary tract infections, and tuberculosis.

Public health experts have been calling on meat companies to cut back on antibiotic use, and on the FDA to enact stricter regulations on the issue like its European counterparts have done. I predicted neither the FDA’s modest actions nor industry’s voluntary agreements over the past few years would have made a difference in cutting antibiotic usage in the meat business, and I was right.

Earlier this month, the FDA released data that showed in 2022 there was a 4 percent increase in sales of medically important antibiotics to the livestock sector. To the FDA’s credit, regulations it passed in the mid-2010s did help bring sales down for a couple years, but they’ve been ticking back up every year since 2017. —KT

Entertainment and culture

Top Gun: Maverick will not win Best Picture (75 percent) — RIGHT

Pfft, come on! Yes, the Academy of Motion Picture Arts and Sciences did deign to gift a Best Picture nomination to this paean to the American military-industrial complex and the medical wonders of Tom Cruise’s anti-aging regimen. But let’s be real — the mysteriously unnamed foreign adversary in the film was more likely to blast Pete “Maverick” Mitchell out of the sky than give the Best Picture to an action movie that, arguably, saved movies as an industry coming out of Covid. Seventy-five percent certainty was, in retrospect, way too low.

After all, the last time the Academy gave the Best Picture statuette to a mega-popular action film was 2004’s Lord of the Rings: The Return of the King, but that was a) the third entry in a beloved trilogy of a beloved book series that had long been considered unfilmable, and b) had elves. Neither was true of Cruise’s air-combat masterpiece, which, besides pulling in some $1.5 billion at the box office, definitively proved the superiority of sexagenerian human pilots over remote-controlled drones. Instead, the Academy honored the multiversal extravaganza Everything Everywhere All at Once, which means there is now at least one Best Picture-winning film that features a mystical bagel as its central point. (Michelle Yeoh and Ke Huy Quan rocked, though.)

So congratulations, Academy, Tom Cruise basically gave you back your business, and for the fourth time, you sent him home with nothing. May you be haunted by whatever weird classical music ghost was poltergeisting around Lydia Tár’s palatial Berlin flat. —BW

The Philadelphia Eagles will win the 2023 Super Bowl (25 percent) — WRONG

Okay, so, I see it says here that the Kansas City Chiefs defeated the Philadelphia Eagles 38-35 in Super Bowl LVII in what was apparently the third most watched television program of all time. Weird — despite being a lifelong Eagles fan, I have no memory of this. But if I had watched the game, I would probably point out:

  • The field was suspiciously slippery in a way that neutralized the Eagles all-world pass rush against Chiefs QB Patrick Mahomes, who is apparently exempt from the laws of physics — and all the NFL could do was blame Rihanna’s halftime show.
  • What kind of treatment allowed Mahomes to do plays like this on what was supposedly a very serious ankle sprain, and was it covered by State Farm?
  • How does a quarterback like the Eagles’ Jalen Hurts just… drop the ball like this? Not dropping the ball is an important part of being a quarterback, given that it is very hard to play offense if you do not, in fact, hold the ball in your hands.
  • At halftime of Super Bowl LVII, the Eagles held a 10-point lead, which gave them an implied win probability of 82.1 percent — significantly higher, I would point out, than the 25 percent I gave in my prediction. (For what it’s worth, the Eagles also had an 82.1 percent chance of winning their division this year before playing the Dallas Cowboys on December 10, who then soundly thrashed them 33-13.)
  • And most importantly… HOW DO YOU CALL A HOLDING PENALTY LIKE THIS ON THE KEY PLAY OF THE MOST IMPORTANT SPORTING EVENT IN OUR (well, mine, at least) LIFETIME?

The lesson here is clear: Don’t put probabilities on your dreams. And maybe don’t drop the football. —BW

Future Perfect

Black Nazis? A woman pope? That’s just the start of Google’s AI problem.

Future Perfect

Should Big Pharma pay poor countries for finding new diseases?

Future Perfect

9 charts that show US factory farming is even bigger than you realize

View all stories in Future Perfect

Sign up for the newsletter Today, Explained

Understand the world with a daily explainer plus the most compelling stories of the day.