Money Stuff: Robinhood Adds Some Agents

For a while now, I have argued that “eventually, every possible trade will be packaged into an exchange-traded fund.” Brokers and wealth advisers have a long history of coming up with complicated ideas to pitch to their more adventurous clients, but ETFs allow brokers to put quite complicated pitches into something you can buy with a click of a button. Ideas like “buy stock and sell call options on it” or “borrow money to buy more stock” or “borrow money to buy Bitcoin and gold” or “long Nvidia short Intel” or “sell crash insurance” are all trades that now exist in ETF form. There are other ways to do these trades — structured notes, or just doing the individual component trades separately — but the ETF is a clean simple package with a catchy ticker that you can buy in your Robinhood account.

But now I think that this argument was naïve, and the ETF, as a universal package for trade ideas, might already be outdated. “Every possible trade” will be an ETF? There are infinite possible trades, some of which ETF providers and brokers and wealth advisers have probably never thought of. If in your shameful secret heart you want to do a trade like “buy every stock whose ticker begins with C, and sell out-of-the-money call options on every stock whose ticker begins with D,” nobody will sell you an ETF like that and you’d be too embarrassed to ask. 

What you want is a more general-purpose technology, a way to describe any imaginable trade in a few sentences and then execute it with a click of a button. And that, increasingly, exists. Bloomberg’s Paige Smith reports:

Robinhood Markets Inc. customers will soon be able to direct artificial intelligence agents to trade equities and make purchases on their credit cards for them.

A client will able to establish an agentic trading account that’s entirely separate from the customer’s standard portfolio, limiting the funds an agent can access to only those deposited, according to a statement Wednesday. The investor can then direct the AI agent to build a diversified portfolio from scratch, for example, or adjust concentrations to take advantage of opportunities as they arise.

“Some people are setting up hypotheses, some people are doing rebalancing,” Abhishek Fatehpuria, Robinhood’s vice president of product management for brokerage, said in an interview. “When we enable customers to connect their outside agents and outside tools, the possibilities are pretty vast.”

Fatehpuria also said he’s looking forward to learning how different people use the product, which will inform how the team develops it.

Here’s Robinhood’s announcement, with some examples of the possibilities:

A long-term investor can have their agent analyze their portfolio for concentration risk and sector exposure, identify where they're over or underweight, and execute a rebalance.

A thematic investor with conviction in AI or semiconductors can have an agent build an initial portfolio matching their criteria, monitor the space for new entrants and analyst upgrades, and rebalance toward the strongest opportunities at a regular cadence.

An active trader can backtest a mean reversion strategy to see how it performed historically, and deploy it to automatically buy oversold stocks and sell when they revert to the mean.

Is there an ETF of “semiconductor stocks that are good”? Sure. There’s even a mean reversion ETF. But you might quibble with their methodologies or rebalancing frequencies; you might want a slightly different buy-the-best-semiconductor-stocks or mean-reversion trade. You can type whatever you want into the box, and the AI agent will give you that. Why use someone else’s off-the-shelf set of criteria for picking semiconductor stocks, when you can roll your own with the help of a friendly robot?

A couple of points. First, this is really a combination of two modern technological developments. One is agentic artificial intelligence, but the other is free stock trading. An agent that regularly rebalanced your $10,000 stock portfolio, never mind one that was constantly buying oversold stocks and selling them when they reverted to the mean, would have no appeal if you had to pay $9.95 per trade. Free retail stock trading — which didn’t really exist a decade ago — enables much more trading, which is a difference in kind, not just degree. In a world of expensive trading, most people’s investment theses have to be, essentially, stocks: ”This company is good, so I will buy its stock and hold it for a while.” In a world of free trading, your thesis can be something more general and thematic — “buy stocks that went down yesterday” or “buy stocks that went up yesterday” or “buy stocks that were mentioned on television this morning” — and you don’t have to worry about the details of implementing it with stock trades. If it requires making dozens of trades per day, that’s no problem.

Second, and relatedly: Stock trades are not really free, [1]  and anyone who works in professional quantitative trading was horrified by that last paragraph. Presumably these agents will be good for business at the electronic market makers who fill Robinhood’s orders and collect a spread on each trade. “Agentic trading,” to them, means something like “retail traders making lots of small predictable rule-based stock trades all the time,” which is what they want. And while “Agentic Trading is launching in beta with support for equities only,” Robinhood plans to add “support for options, crypto, event contracts, futures, and more.” There are few more pleasing phrases, to a market maker, than “automated retail options trades.”

Not that long ago, I assumed that the future of retail investing was ever-increasing levels of boring passive indexing, not only because that is sensible (it’s hard to beat the market in your free time) but also because it is good customer service: Stuff should be simple and easy, and indexing is simple and easy. But in fact those features are not intrinsically linked, and in the modern world an automated zero-day-options-YOLO’ing strategy will not be significantly more difficult to implement than passive indexing. Or an automated sports-betting strategy, for that matter.

Third, I really have no idea how popular these tools will be. (I’m perhaps exaggerating how user-friendly this is; Robinhood’s overview says “You’ll first need to connect a third-party AI agent and then follow the on-screen steps to connect it with your Agentic account through the Robinhood Trading MCP (Model Context Protocol),” so it’s not like you just type your instructions in a box and a Robinhood robot carries them out.) We talked last month about a similar agentic AI push at Public, another retail brokerage, and I wrote:

We have talked a few times before about the possibilities for AI to coordinate retail behavior. If retail investors tend to get investment ideas from AI chatbots, I have written, those chatbots might tend to send all the retail investors to the same ideas, so the stocks favored by the chatbots will go up. Similarly, here, if retail investing becomes increasingly agentic, the agents might tend to automate stereotypical retail behaviors. These days, the most stereotypical retail behavior probably is buying the dip, and maybe that is the first one that will be automated. And then if the S&P 500 drops 2% or more in a trading session, every hedge fund will buy at the close, knowing that there’s billions of dollars of agentic retail demand coming the next morning.

One possibility is that widespread agentic retail investing would allow each individual investor to automate his or her idiosyncratic investing ideas, leading to, you know, a lot of random noise trading. Another possibility is that widespread agentic retail investing would lead to all the individual investors automating the same, like, 20 investing ideas, leading to herding and predictability.

Anyway Robinhood also released an agentic credit card, with its own list of possible trades:

A sneakerhead can tell their agent to buy a coveted new release in their size whenever it drops below $300.

A foodie can instruct their agent to book the most exclusive restaurant reservation in town as soon as their preferred date and time becomes available.

A small business owner can use their agent to fill multi-item order lists (e.g. buy me the ingredients to make a custom cake without spending more than $50), or even buy a website domain.

It’s a little annoying that they’re separate. Surely the final state of all of this is, like, “build a relative valuation model that tells me when semiconductor stocks, sneakers, Knicks bets or 4 Charles reservations are below fair value, and buy whatever the best deal is.” 

The basic theory of the golden parachute is that it is often good for shareholders if their company is acquired by a bigger company at a premium, but it is generally bad for the chief executive officer. The CEO has a nice job that pays her a lot of money and makes her feel important; if she sells her company, she will probably lose her job. Therefore, public companies often have generous golden parachutes — change-of-control severance arrangements for their top executives — to encourage the CEO to sell if that’s what’s best for shareholders. 

There are various flaws in this theory. Here’s one: If the CEO is about to get fired anyway, the golden parachute gives her an incentive to sell the company, even if that’s bad for shareholders. If she sells, she loses her job and gets a big payout. If she doesn’t sell, she loses her job anyway and gets a much smaller payout, or possibly no payout if she has done something bad enough to get fired “for cause.”

That is: Usually it is good for shareholders to nudge the CEO toward selling the company, but sometimes that nudge might be too much. “Well things are hopeless here for me,” the CEO might conclude, “but at least I can sell the company and cash in.”

Is that a real thing? I kind of doubt it, but occasionally you get a fact pattern that is good enough to give it a go. Bloomberg Law reported last week:

Microsoft Corp. agreed Friday to pay $250 million to end shareholder litigation over its $75.4 billion acquisition of Activision Blizzard Inc., according to court filings.

The agreement, filed in Delaware’s Chancery Court, would resolve claims that former Activision chief executive Bobby Kotick and other board members signed off on a heavily discounted transaction to get out from under the sexual harassment crisis engulfing the gaming behemoth. A judge let the case move forward against them last October, though she rejected allegations that Microsoft exploited the scandal by colluding with them to drive down the deal price.

The lawsuit, filed by a Swedish pension fund, alleged that the ex-CEO—facing a thicket of damaging media coverage and litigation on all fronts—tipped Microsoft about his desperation to get a deal done, giving the tech giant the chance to orchestrate a lowball offer.

Here is the complaint, and here’s Kotick’s answer. There are various problems with the lawsuit, including that (1) no agency ever found Kotick responsible for Activision’s sexual harassment problems, (2) Kotick was a big shareholder and doesn’t even seem to have had a golden parachute (the theory is that he was trying to hang on to unvested options that he’d lose if terminated for cause) and (3) actually the price was pretty good? (“Today, given that console sales are at an all-time low and Call of Duty sales are off over 60 percent from the prior year, Plaintiff should be expressing extreme gratitude for the foresight Activision leadership demonstrated in consummating this transaction,” says Kotick’s filing.) But the basic theory hangs together, it was worth a shot, and apparently it paid out.

I know this makes people mad, but really, if you have an index that is meant to reflect substantially all of the US stock market, then it will have to include SpaceX when SpaceX goes public. Will SpaceX be part of the US stock market? Yes. Will it be a big part? Seems so! Therefore, the index has to include SpaceX. Maybe not every index: The Dow Jones Industrial Average is small and weird, and even the S&P 500 has some history of requiring seasoning and profitability that might exclude SpaceX, though don’t bet on it. But the Russell 3000? Sure sure sure, you gotta wave it in:

FTSE Russell adopted a rule change that will speed the addition of newly listed large-cap companies to its main indexes, weeks ahead of SpaceX’s expected record-breaking initial public offering.

Under the policy announced late Tuesday, IPOs with investable market capitalizations above the cutoff for the Russell Top 500 will qualify for so-called fast entry after their fifth trading day. Previously, eligible companies were assessed during quarterly reviews.

“The introduction of a fast entry mechanism for sizable IPOs enables the indexes to reflect significant market developments more promptly,” Arne Noack, head of equity and multi-asset indices, Americas at FTSE Russell, wrote in a statement.

The index provider, owned by London Stock Exchange Group Plc, is following a similar move by Nasdaq Inc. earlier this year aimed at accommodating potential megadebuts. The changes highlight pressure on index providers to adapt as companies stay private longer and go public at much larger valuations. Nasdaq — the expected listing venue for SpaceX — already shortened the waiting period for index inclusion to 15 days from at least three months. S&P Dow Jones Indices is still evaluating possible revisions.

More than $30 trillion in assets globally are benchmarked to indexes whose inclusion rules are either already in effect or under review. The push for faster inclusion has raised concerns among some investors, who say adding IPOs too quickly could expose passive funds to greater volatility and force them to buy shares before reliable market pricing is fully established.

Here is FTSE Russell’s announcement. While the index obviously has to include SpaceX, I am not sure when it has to do so. S&P is considering going from 12 months to six, which seems pretty sedate; Nasdaq’s 15 days seems reasonable; FTSE Russell’s five days is maybe a little tight? Index funds tend not to actually buy shares in the IPO, because the stock is not generally in the index at the time of its IPO, but at five trading days, will index arbitrage desks of hedge funds be buying in the SpaceX IPO? At what price? The theory of an index fund is that active investors figure out what the stock prices should be, and the index fund just copies their work, but if index funds are more or less buying in the IPO then they are helping to set the price themselves.

If you are an American company that sells brake parts in the US, and you have a factory in China that manufactures the brake parts, and the factory in China charges your American operation for the brake parts, how much should it charge? The two basic answers are:

1. Whatever the going market price of the brake parts is, or
2. Whatever minimizes your taxes.

In simple terms, if the US charges tariffs on brake-part imports from China, and if those tariffs are a percentage of the purchase price, you should charge approximately $0 for the brake parts, because then you will pay $0 in tariffs. [2] Everybody knows this, so it is not allowed. Instead, your Chinese subsidiary is supposed to charge you the market price for the brake parts. Are there incentive and opportunity to shade this a little bit? Oh sure. Bloomberg’s Steven Church and Jonathan Randles report:

The US government has joined the line of creditors impacted by alleged fraud at bankrupt auto-parts maker First Brands, in this instance for accusations the company cheated on how much it should have paid on tariffs.

First Brands was hit with a formal claim for $285.5 million related to allegations that it underpaid levies on parts imported from China. The figure includes a demand for penalties as well as the unpaid tariffs. ...

The US claim is based on a whistleblower complaint filed in March 2022 in federal court in New York and made public earlier this year. That complaint was filed by Alder Wood LLC, which argues that the company imported automotive brake parts from its subsidiary in China without paying the proper amount of tariffs.

Here is the claim, which attaches the whistleblower complaint. It is pretty straightforward. First Brands has some US subsidiaries called Brake Parts Inc. and Centric Parts; Brake Parts Inc. has a Chinese subsidiary called Longkou Haimeng Machinery Co. Ltd., which makes its brake parts; First Brands decides what prices Haimeng will charge it for its parts; and “those prices — after having undergone a series of arbitrary reductions at the direction of First Brands — are only a fraction of arm’s length prices.” Allegedly this is pretty observable: “In the automotive brake device industry, various manufacturers offer identical products manufactured to the same specifications and bearing the same parts numbers. These products are essentially fungible commodities.” First Brands buys identical parts from Haimeng and from other unrelated manufacturers, “as part of an effort to diversify its product sourcing beyond Haimeng, which has limited capacity,” and allegedly pays about twice as much to third parties as it pays to Haimeng.

Also:

In or about January 2022, BPI paid Haimeng an additional $40 million, apparently once it became clear that the price reductions ordered by First Brands had left Haimeng financially strapped.

Specifically, the $40 million payment came about after Chinese authorities raised questions about Haimeng’s 2021 financial results. Given the large year-over-year decline in Haimeng’s revenues (resulting from the transfer price reductions), those authorities likely suspected that Haimeng was concealing revenues to evade taxes. …

In fact, BPI employees have internally dubbed the payment a “make up payment.”

Yes if your supplier sells you parts for less than it costs to make them, your supplier will eventually run into financial troubles. But if you occasionally throw in an unrelated Christmas present, maybe that’s fine.

At the New York Times Magazine, Adam Iscoe has a delightful story about pseudonymous sharp bettors who make money on Kalshi and Polymarket. The highlight might be this guy:

@PrinceHal, a struggling screenwriter turned full-time Kalshi trader, has been trading for about a decade. He showed me how he builds inflation-forecasting models that consistently outperform major financial institutions to the tune of $3.7 million in lifetime profits. “Me and the banks are doing the same thing,” PrinceHal said. “I think it says more about the market than it does about me. They can’t beat an acting major in a garage with Excel.” (I asked a macro-economist who forecasts inflation for a prominent Wall Street hedge fund to review PrinceHal’s trades and methodology. “I’m actually dumbfounded,” he told me. “This guy might literally be Nostradamus.”)

Yeah man if a “prominent Wall Street hedge fund” says you “might literally be Nostradamus,” you are selling yourself short at $3.7 million in lifetime profits. I also enjoyed the contrast in these two quotes:

A Kalshi spokesperson told me, “Traders tell us that Kalshi is fairer and less predatory than casinos and sportsbooks because there is no house that wins when customers lose.”

And:

“There’s a lot of sharps out there, and they’re winning — and guess who they’re beating?” [Susquehanna International Group co-founder Jeff] Yass said. “Us,” he said. “We’re a big share of the volume, so when they figure something out, and they want to make a sharp bet, we’re often on the other side of it.”

“There’s no house,” says Kalshi. “Lol we’re the house,” says Susquehanna. I wonder who’s right?

Why Spirit Airlines Failed. How AI threatens the giants of consulting. BP’s Ousted Chair Says He Was Fired Without Explanation. Inside BP’s boardroom ‘bullying’ scandal. Memory Chip Frenzy Sends SK Hynix, Micron Into $1 Trillion Club. The world’s biggest EV maker weans itself off supply-chain finance. US Says $20.6 Billion of Tariff Refunds On the Way to Importers. Insurance Mogul Lindberg Gets 12 Years for $2 Billion Fraud. For Wall Street’s Private Investments, In-House Insurers Are the Go-To Buyer. Polymarket, Kalshi Face Spanish Ban. Ford’s Stock Is Surging—and It’s Got Nothing to Do With Its Car Business. EU defence chief urges states to stop making ‘haute couture’ missiles. “Promised amenities, including a restaurant bunker, a pool bunker and a horse-stable bunker, have yet to materialize.” 

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks!