Thursday links: financial nihilism, prediction market ETFs, non-wisdom of crowds, AI trading agents

Gen Z “investing”

A survey by Northwestern Mutual reports that 32% of Americans between the ages of 18 and 29 say they are either investing in or considering investing in sports betting and prediction markets.

The report correctly labels this financial nihilism: sports betting and prediction markets are negative-sum activities and therefore not a form of investing (however you might define it).

It also asked why they would “invest” this way. Of the Gen Z respondents considering sports or prediction market betting, 80% said it’s because they feel they’re falling behind financially. They see sports betting and prediction markets as a way to catch up.

Feelings aren’t facts, however, and the data tell a different story.

Economist Jeremy Horpedahl writes that, measured by wealth accumulation, the combined Millennial/Gen Z cohort is “well ahead of where Gen X was even in their late 30s, and ahead of Boomers at around age 37 as well.” 

Gen Z, in fact, is off to a pretty good start.

But too much sports and prediction market betting might change that.

Event contracts for your 401k

Prediction market ETFs are expected to be approved by the SEC next week.

Yes, that sentence reads like a finance-themed Mad Libs — “prediction market” and “ETF” seem as unlikely a pairing as Jon Bogle and Jesse Livermore — but you really will be able to have your index funds and event contracts listed side by side in a brokerage account.

The first listings are kind of reasonable, though, as they’re all related to politics: Democratic President ETF, Republican President ETF, Democratic Senate ETF, etc.

One could argue these belong on stock markets because investors could use them to hedge their portfolios ahead of elections. I’m sure a few people occasionally will.

But how long will it be until we get a Super Bowl Gatorade Color ETF, Celebrity Breakup by 2027 ETF, and Zombie Apocalypse by 2030 ETF?

Not long, I bet. (If there were a prediction market on that, I’d definitely buy it for my 401k.)

“The line between gambling and investing is artificial and thin,” Michael Lewis wrote in The Big Short.  

It’s getting thinner.

The folly of the crowd

Kalshi says prediction markets are based on the idea that “the wisdom of the crowds can produce truth among noise in the world.”

A study of millions of Polymarket trades, however, suggests the opposite: the uncanny precision of prediction markets is attributable to just 3% of all bettors.

The onchain researchers find that the trades of this elite 3% “predict future prices and final outcomes, make prices more accurate throughout a market’s lifespan, and react to news the moment it arrives,” the study concludes.

The other 97% of accounts are there simply to provide liquidity for them — we are the noise Kalshi mentions.

Well, slightly less than 97%.

The researchers also categorized 1,950 accounts as insiders: accounts that open, make one bet, win, and then stop betting.

These insider accounts earn $15,000 on average — far more than the $3,255 earned by the 3% of skilled accounts. But the study finds there aren’t enough insiders to move the needle on overall accuracy.

“Prediction market accuracy thus reflects the wisdom of an informed minority,” they conclude, “not the wisdom of crowds.”

Too much AI breaks the market

It stands to reason that a market full of hyper-intelligent AI agents would be hyper-efficient.

But a new study finds that when AIs trade, "individual rationality leads to collective fragility."

The risk stems from the reinforcement learning trading agents do — continuously adapting their behavior according to their results.

Because these learning agents are all trained on the same data, they converge on similar trading strategies.

With enough of these like-minded agents put into action, the study warns that their reactions to market events could create momentum effects, causing the market to overshoot in whichever direction it’s going (down, probably).

Worryingly, this sounds a lot like the Black Monday crash of 1987.

Then, futures-trading algorithms — all responding to the same market signals in the same way —  all starting selling at the same time. The selling begat more futures-driven selling and what should have been a run-of-the-mill market correction suddenly threatened to become a systemic collapse: The Dow Jones Industrial Index fell a now-unthinkable 22.6% in a single day.

Gen Z investors don’t know how good they’ve got it.

— Byron Gilliam