Back in March, I laid out my theory of pensions. There are two main components to this theory. One is that people in the US used to have defined-benefit pensions, and now they mostly have defined-contribution retirement plans like 401(k)s. A pension is a big pool of money, managed by professionals, that has to pay thousands of beneficiaries some promised level of benefits every month when they retire. Because the payouts are pretty predictable, the pension fund can make very long-term investments: Some of its money definitely won’t be needed in for the next 30 years, so it can lock that money up for decades. So pension funds can make illiquid investments, and get paid an illiquidity premium: By locking up their money in bespoke, non-traded stuff, they can hope to earn a higher return than they’d get by just investing in Treasury bonds or even the stock market. 

Meanwhile a 401(k) is an individual account; you put money into the 401(k), you pick how it’s invested, and at retirement you can spend it. In the abstract, a 401(k) has similar goals to a pension, and could do similar things: If you are a 35-year-old saving for retirement, you shouldn’t need your money for 30 years, so you can also buy illiquid investments and earn an illiquidity premium. But because a 401(k) is, in most respects, an individual account, it is not so simple to lock it up for 30 years. You probably won’t withdraw money until you retire, but you might. What if you have a medical emergency and need the money? What if you die? A pension fund can smooth out those risks; a 401(k) can’t. 

Therefore, in practice, investments in 401(k) plans need to offer some liquidity. You need to be able to take your money out; mass marketed 401(k) investments can’t have the form “you give us money and we give it back in 10 years, no earlier.” So it is much harder for a 401(k) plan to invest in bespoke, non-traded stuff. It’s not impossible: There are various structures (interval funds, mixed public/private funds, etc.) that are designed to, essentially, do the pooling. If you raise money from retail investors and invest it in illiquid private credit, but you let the investors take up to 5% of their money out each quarter, that’s mostly fine: Fewer than 5% of your investors will probably need money (for emergencies, etc.) in any given quarter, so you will be able to keep your money invested in the illiquid private stuff.

But, if you’re offering liquidity, eventually your retail investors will use it in a correlated way. We talked a lot earlier this year about a sort of quasi-run on retail private credit funds: Those funds often let investors withdraw 5% of their money each quarter, but there was a wave of investor worry about private credit, and investors started asking for 15% or 20% or 40% of their money back. Those investors didn’t all have emergencies; they just wanted different investments. The funds’ underlying investments were illiquid, premised on the idea that they’d lock their money up for a long time and earn an illiquidity premium. But their investors expected liquidity, and were annoyed that they couldn’t get it. After the experience of this year, these funds do not look like a particularly great way to invest retirement savings in private credit. Whereas traditional institutional private credit funds, which raise locked-up money from, you know, pension funds, are doing fine.

The other component of the theory is that, historically, the financial industry could charge people on the order of 1% per year to invest their money in stocks. You’d buy a stock mutual fund in your 401(k), and it would have a 1% management fee. These days, with the rise of index funds, the going rate to invest people’s money in stocks is on the order of a couple of basis points. That is much, much less; it is an existential crisis for a lot of companies that are in the business of investing people’s money in stocks.

And so a critical goal of the financial industry in 2026 is to invest more of people’s retirement savings in “private assets,” because it is still possible to charge fees on the order of 1% (or higher!) for private investments. Private investments are, in essence, not indexable; an asset manager who invests in private assets is more obviously earning her keep than one who invests in public stocks. 

When you combine those two components, the conclusions are:

1. The investment industry wants to invest more retirement savings in private assets, but
2. The 401(k) is a bad vehicle for investing in private assets, and creates a lot of liquidity risk for investment firms, so the investment industry would prefer for more of people’s retirement savings to be in the form of pension funds.

The first point is, like, super obviously true; we have talked a lot about the push to put private credit and private equity investments into 401(k) plans. (Also crypto and sports gambling, but that’s a different story.) 

The second point is less obvious: Historically, the move from defined-benefit pensions to defined-contribution 401(k) plans was viewed as a big win for the financial industry. But that was back when running a stock mutual fund was lucrative.

Anyway I wrote in March:

It would be interesting if the story of the next decade in retirement investing is a move back to something like pensions, an increasing emphasis on pooled guaranteed-income vehicles rather than atomistic individual self-managed lump-sum investment accounts, because those vehicles are better for private-market investing, and everyone is solving for private-market investing.

And today the Wall Street Journal reports:

BlackRock will offer American workers a chance to invest more like a multibillion-dollar pension.  

The world’s largest investment firm by assets under management said it would work with corporate clients to build customizable funds for 401(k) plans that can include slices of public stocks and bonds, private assets and even annuities that provide guaranteed income in retirement.

The new framework aims to resemble the menu of options offered to public-employee pension plans, university endowments and other deep-pocketed institutions that hire firms like BlackRock to manage their money. ...

“We’re trying to take the best of what 401(k)s have done, such as portability and ownership by the individual, and combine that with the best of what pensions did, which is professional management, long-term thinking and certainty through income in retirement,” [Nick Nefouse, BlackRock’s global head of retirement solutions] said in an interview. ...

BlackRock, Vanguard, Fidelity Investments and State Street have all recently announced new funds that give investors the option of buying an annuity before or after their retirement date that guarantees some level of income for the rest of their lives.

The latest debate in retirement investing centers on private assets. President Trump signed an executive order in 2025 directing federal agencies and regulators to clear the path for assets such as private equity, private credit and real estate into 401(k) plans.

Private-asset funds have long been a major component of pension-fund and insurance-company portfolios. Proponents of private-market funds in 401(k)s argue that they can be less correlated to public stocks and offer premium returns in exchange for less liquidity. Skeptics point to higher management fees and less transparency from managers. 

I mean, both things are obviously true. The advantage of a pension-fund-like structure, for investors, is that it allows them to earn an illiquidity premium on money that they really should be able to lock up for retirement. The advantage of a pension-fund-like structure, for asset managers, is that it allows them to earn higher fees.

In 2022, an employee at OpenSea insider traded on nonfungible tokens. I realize that sentence probably sounds like nonsense, so a brief refresher. Nonfungible tokens, or NFTs, are crypto tokens representing like pictures of monkeys or whatever, which in 2022 people pretended were valuable. OpenSea was the leading platform for trading NFTs. 

What determined the price of an NFT? NFTs were kinda sorta a form of “art,” and you could imagine the price of an NFT being determined by, like, its aesthetic quality, or its provenance, or how well it fit with the decor of your Twitter profile, or some other art-like characteristics. But in practice, absolutely not.

In fact, what often determined the price of an NFT was whether it was featured on OpenSea’s homepage. There was a vogue for buying NFTs, and so people bought the NFTs that were closest to hand. The NFTs featured on the homepage were the obvious ones to buy, so people bought them and their prices went up. The value of an NFT was determined, at least in part, by how much attention the leading NFT site drew to that NFT. More cynically, the value of an NFT was determined by how easy it would be to sell to a greater fool.

And therefore, the paradigmatic way to “insider trade” NFTs was to be an employee at OpenSea who knew what NFTs would be featured on the homepage. You could buy those NFTs at low prices, when they were the unknown work of struggling undiscovered NFT artists, and then quickly sell them at high prices, when they were featured on the homepage.

And a guy did that, and he was arrested and charged with wire fraud. And he was convicted and sentenced to three months in prison. And then in 2025 a federal appeals court reversed his conviction, finding that actually he didn’t commit any wire fraud. We talked about the decision at the time, and it is quite perplexing, but basically the court found that, while being featured on OpenSea’s homepage was obviously material to the price of an NFT — arguably it was the only thing that was material to the prices of many NFTs — it was not clear that it was material to OpenSea. “The evidence also indicated that the featured NFT information was so tangential to OpenSea’s business that failing to maintain the confidentiality of the featured NFTs would not affect users’ attitudes toward the platform,” said the court. Insider trading, I often say, is not about fairness; it’s about theft. It’s illegal insider trading to steal commercially valuable information from your employer and use it to trade, but if the information “lacked commercial value,” then go ahead and insider trade on it. Not legal advice. That seems to be what the opinion says, but I’m not sure that it’s right.

Anyway now it is 2026 and the main way to insider trade in crypto is, still, to get advance notice that a big platform will list a token. You buy the token, it gets listed on the big platform, now it’s easier to sell, so its price goes up. And so:

Two former Robinhood Markets Inc. employees were charged by US prosecutors with fraud for allegedly using nonpublic information to trade crypto-linked perpetual futures on a decentralized derivatives exchange.

The engineers, identified as Hefu Chai, 36, and Huaisong Xiang, 30, allegedly used confidential information related to whether and when additional cryptocurrencies would be supported by Robinhood to buy crypto-linked perpetual futures on the exchange Hyperliquid, according to a statement Tuesday from Manhattan US Attorney Jamie McDonald. The two staffers allegedly made more than $50,000 each from their trades.

Sure. Here are the press release and the complaints against Chai and Xiang. It’s stuff like this:

On or about December 1, 2025, CHAI received a Slack message on the private Slack channel that Robinhood would launch AERO, a cryptocurrency token, on or about December 4, 2025, at 8:30 a.m. On or about December 4, 2025, CHAI, using Wallet-0x548, opened long positions on Hyperliquid for perpetuals with AERO as the underlying asset. On or about December 4, 2025, Robinhood publicly announced that AERO had been listed on Robinhood for trading. After Robinhood's public announcement, CHAI closed his long position in AERO.

CoinMarketCap tells me that AERO started that day at about $0.683 and peaked at about $0.7094 that afternoon, so the Robinhood listing was maybe not that life-changing, but it’s something.

But … is it illegal? After the OpenSea case, are crypto exchange employees just sort of allowed to do this? Well, in that case, the appeals court’s problem was that the government hadn’t proved, and the jury hadn’t found, that the listing information had “commercial value” to OpenSea. The prosecutors in this case are trying to fix that problem. From the complaints:

Information regarding the digital assets to be listed on Robinhood Crypto was commercially valuable to Robinhood and maintaining the confidentiality of contemplated future listings was important to Robinhood's ultimate commercial objectives. Among other things, Robinhood Crypto competes with other cryptocurrency exchanges on the breadth of its offerings. Premature public disclosure of Robinhood's future listings would allow Robinhood's competitors to pursue similar listings and thus deprive Robinhood of the competitive advantage of listing those assets first. Additionally, because of the size and popularity of Robinhood's platform, the listing of a particular digital asset on Robinhood can create upward price pressure on the asset by generating new demand. The premature disclosure of such listing information can lead to a reduction in Robinhood's revenue because customers who would otherwise trade these digital assets on Robinhood Crypto once listed could instead trade them on a competitor's platform to profit from the expected price increase that would follow Robinhood's listing. Further still, Robinhood assesses its brand and reputation to be important to its ability to attract and retain customers. Premature disclosure of its plans for future listings, especially when that disclosure is selective and unauthorized, risks reputational damage that may diminish Robinhood's standing with current and potential customers.

Insider trading law is so weird because obviously the point here is that the Robinhood listing was material to the tokens’ prices, not that it was material to Robinhood. But because of a weird legal technicality, the prosecutors have to say it was material to Robinhood. Was it? Maybe, I don’t know, it’s not like Robinhood is trading these tokens. “Token prices spike when Robinhood lists them, and if other platforms knew about that in advance then they’d capture the spike, not Robinhood.” Sure I guess.

I realize that this is not the thing that most people find most interesting about Jeffrey Epstein, but to me the most interesting question about Epstein is: Was he a natural genius of tax law? He did not have a law degree and does not seem to have had much formal training in tax planning, but a number of very rich people seem to have paid him enormous amounts of money for tax advice. Given what else Epstein was getting up to, people occasionally wonder if those payments were entirely for tax advice. To resolve that question, it would be helpful to know how good and differentiated his tax advice was.

And I can’t tell! One the one hand, there is no, like, monumental edifice of tax avoidance that bears his name. Nobody is like “oh he invented ETF heartbeat trades” or “oh he invented buying real estate for depreciation deductions” or whatever. In all of the stories I have read about Epstein, I have never read anything of the form “he told Client X to do Strategy Y, which nobody had ever thought of before and which saved $Z of taxes.”

On the other hand, there is a general vague vibe like that. Lawyers for Apollo Global Management concluded that Leon Black paid Epstein oodles of money for tax advice that “conferred more than $1 billion and as much as $2 billion or more” in tax savings, so, sure, good. I appear in the Epstein files because I wrote a column about a fairly clever trade that seems to have saved about a billion dollars in estate taxes. Was this trade a work of unprecedented genius? No, but it was pretty cute, and whoever came up with it probably got a big fee and deserved it. Was this trade a work of Jeffrey Epstein? Almost certainly not! But a fancy estate-planning lawyer at Paul Weiss did email Epstein my column with a note saying “I thought of you when I read this article. Was this your idea?” The point is that fancy tax lawyers, upon seeing fancy tax trades, thought “ooh this looks like an Epstein.” Which suggests that the fancy tax lawyers had previously seen some actual Epstein trades and found them impressive.

But there is contrary evidence. Earlier this month, Bloomberg News published an investigation of Epstein’s tax advice in which outside experts were unimpressed:

“On the surface, the only remarkable aspect of this planning is the price tag,” said Victoria J. Haneman, a University of Georgia law professor specializing in tax issues and estate planning. “You could probably employ all of the top law firms in New York at the same time on the same estate plan and not hit $150 million” in fees. …

Many of the tools Epstein employed in the tax transactions cited in the Dechert report have been used by other billionaires to generate massive tax savings. Those include using grantor-retained annuity trusts, or GRATs, a specialized type of time-limited trust which can allow individuals to bypass estate taxes; leveraging the tax code to defer capital gains taxes on art deals; and splitting assets into multiple holding vehicles to enjoy discounts to their taxable values.

Today Bloomberg’s Tom Schoenberg and Jeff Kao have another story about how law firms would retain Epstein as a consultant, for token fees, in order to bring him under the umbrella of attorney-client privilege:

By hiring Epstein as an outside expert, the firm effectively was handing him the shield of a licensed lawyer. Certain communications — in-person meetings, phone calls, emails, texts — would remain strictly confidential. …

Protected by attorney-client privilege, Epstein advised lawyers on issues involving Black, co-founder of Apollo Global Management. Meanwhile, Epstein referred rich and influential figures within his network as potential clients, the documents show. …

Known as Kovel letters, after a landmark case from 1961, these arrangements extend attorney-client privilege to people who aren’t lawyers. The outsiders typically are accountants or other financial professionals, but not always. Lawyers in the Martha Stewart insider-trading case, for example, extended legal privilege to a public-relations firm.

This looks suspicious, if your model of Epstein is “mostly sex offender,” but if your model of Epstein is “he was the Michelangelo of tax minimization but did not actually have a law degree,” then you could see why law firms would want to work with him to advise clients on tax planning, and to have his advice protected by attorney-client privilege. But Schoenberg and Kao quote some skeptics:

Bruce Green, director of the Louis Stein Center for Law and Ethics at Fordham School of Law in New York, questioned why Epstein would be considered crucial to a lawyer’s work in the first place.

“Is he the foremost expert on financial matters and tax that are going to be useful to the estate planning,” Green said. “It’s hard to imagine.”

Is it? 

We talk from time to time about how the financial-industry recruiting pipeline keeps starting earlier. When I was a lad, you could become an investment banker by stumbling into a recruiting session in the spring of your senior year of college, and you could become a private equity associate by completing a two-year analyst program at a bank, doing some private equity deals, thinking “that looks cool” and sending in a resume. In recent  years, though, it has become normal to get your private equity job before starting at your banking job (though there has been recent pushback). And investment banking recruiting now depends on membership in competitive college finance clubs that effectively require candidates to be expert financial modelers by the time they graduate from high school. It’s all a little grim. “Eventually you will put your child’s name on the KKR waiting list the moment she is conceived,” I joked last year.

Here’s a Bloomberg Businessweek story about “What It Takes to Get Your First Job on Wall Street,” which lays out the grim realities. But it also features this astounding sentence:

One college junior in New York learned about the industry when his fifth grade teacher told him he’d make a terrific investment banker.

He’s interning at Evercore next summer, so I guess the teacher was onto something, but: What? What? I happen to know some fifth-graders, and statistically it is likely that one or two of them will grow up to be terrific investment bankers, but I would never want to guess which ones. [1] What was the tell? “He was building discounted cash flow models at recess,” har har har, but obviously not, right? He had a firm handshake and a good golf game? He was decent at math and all the other kids came to him for advice on Pokémon trading strategy? His father was the chief executive officer of a big public company? 

OpenAI Weighs Funding Round at Over $1.2 Trillion Valuation. ‘Like It’s Raining Gold’: The AI Boom’s Unexpected Billionaires. AI bosses’ safety push sparks rift inside OpenAI and Anthropic. Nvidia and Meta bosses reject efforts to co-ordinate AI slowdown.  Crypto Stocks and Tokens Drop After Senate Blocks Landmark Bill. Hedge Funds Are the Wild Card in the Turbulent Bond Market. What’s Inside the Bond Market’s ‘ Toxic Stew.’ Hedge fund warned ousted lawyer to settle £36bn BHP dam collapse claim. Glencore Sued For $2 Billion; Accuses Radiant World of Fraud. Hackers say they breached Italian state email to target Revolut ‘crypto whales.’ The Stagnant Housing Market Is About to Face a 7% Mortgage. Oil Executives Say the Great Fuel Crisis Is Here. “After his indictment, Roberts closed his art gallery and briefly operated an outlet down the street selling Labubus.”

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!