Money Stuff: Tax Alpha Needs Some Real Alpha

We talked last month about tax alpha. The basic idea is that it is hard to beat the market by buying the stocks that go up and avoiding the stocks that go down. It is, however, relatively easy to buy some stocks that go up and other stocks that go down. Owning stocks that go up is good (you have more money), and owning stocks that go down is also good: You sell them, realize a tax loss, and use that loss to offset other gains and reduce your taxes. [1]

This is pretty straightforward; it is called “tax-loss harvesting” and lots of financial advisers and robo-advisers will do it for  you. But some money managers will scale it up for you. You can borrow money to buy more stocks, generating both more gains and more tax losses. This increases your market risk — owning $200 of stocks with leverage is riskier than owning $100 of stocks without leverage, even after the tax benefits — but you can offset that risk, and generate even more tax benefits, by shorting some stocks. Put in $100, buy $200 of stocks, short $100 of stocks: Now you are net long $100 of stocks. Some of your longs will go up and some of your shorts will go down and you will make money; other longs will go down and shorts will go up and you will generate tax losses. [2] Overall, if you are long and short reasonably well-constructed diversified baskets of stocks (say, randomly sampling 200 long stocks and 100 short stocks from the index), you should expect to generate roughly the market return on $100 of stocks, but with some extra tax juice.

Now, there is a long history of hedge funds — “long/short equity funds” — that take $100, buy $200 of stocks and short $100 of other stocks, for a $100 net long position. These hedge funds traditionally do not care about harvesting tax losses. For one thing, many of their clients are tax-exempt and pay them only for pretax profits; for another thing, they are trying not to have losses. The goal of a traditional long/short equity fund manager is to (1) buy $200 of stocks that go up and (2) short $100 of stocks that go down. Obviously this is hard — harder than generating the market return — but that’s why hedge fund managers, traditionally, get paid the big bucks. These hedge funds are charging for alpha, for market-beating returns.

My point was that beating the market — generating pre-tax alpha — is pretty hard, because you have to pick the stocks that will go up, while generating tax savings is pretty easy, because any diversified basket of stocks will have winners and losers and you can gross it up mechanically to generate a lot of tax losses while still earning roughly a market return.

And so tax alpha should be a much more mass-market product than pre-tax alpha. Instead of being a rare and mysterious thing that only special geniuses can produce in invitation-only hedge funds that charge high fees, it could be something that any reasonably sophisticated asset manager or robo-adviser could roll out broadly to customers for modest fees. Hedge fund strategies that generate pre-tax alpha plausibly have limited capacity; the strategy of “gross up the market return and harvest tax losses” does not.

When we talked about this last month, I quoted a Bloomberg News story about tax-aware offerings from AQR Capital Management, Two Sigma and other hedge fund managers, which included this paragraph:

“Tax-aware strategies must be built on credible pre-tax alpha and that should be the primary draw for investors,” an AQR spokesperson says in an email. “But delivering strong after-tax outcomes requires more than just time-tested alpha — it demands sophisticated tax-aware resources and infrastructure, which we’ve built into every facet of AQR.”

And I wrote:

I feel like the obvious lesson here is that tax-aware strategies don’t have to be built on credible pre-tax alpha? I mean, pre-tax alpha is nice, but getting the broad market index return and not paying taxes on it would be great, probably better than the alpha most hedge funds can achieve, and also probably easier.

I got several emails to the effect of: “Shh, you’re not allowed to say that.” [3] The US Internal Revenue Service, and the courts, take a dim view of trades that are done only for tax benefits; you need to have some “economic substance,” some non-tax purpose for your transaction, to get the tax benefits. The exact trade that I have described here — (1) put in $100, (2) buy $200 of random stocks targeting the index return, (3) short $100 of random stocks targeting the (negative) index return, (4) overall earn the index return on $100 of net stock exposure and (5) harvest tax losses on the $300 of gross exposure — might be too cute. If you did that, and certainly if you explicitly advertised that you were doing that, then the IRS would come after you, your customers would lose the advertised tax benefits, and you would get in trouble.

The takeaway is:

1. The people best positioned to offer tax alpha strategies are people — AQR, Two Sigma, etc. — who already run hedge funds, and who have some credible basis for saying “actually we generate pre-tax alpha by going long a bunch of (mostly good) stocks and going short a bunch of other (mostly bad) stocks, though of course sometimes our longs go down and then we harvest the tax losses.”
2. They definitely have to say that. As opposed to just “we grossed up the index return to get tax losses.” “Tax-aware strategies must be built on credible pre-tax alpha,” and say so, to placate the IRS.

Anyway here’s a Financial Times article about how these strategies are all the rage:

“It’s the hottest product in Greenwich,” said one person who works on Wall Street of tax-aware funds.

Fathers at a recent local sports game were fixated on AQR’s Delphi Plus fund, which specialises in ordinary income losses, the person added. “Everyone’s piling into them now.”

(The trades that I have schematically described here generate only capital losses, which can mostly only offset capital gains, but “AQR and Quantinno differ from earlier iterations of tax-aware investing strategies by seeking to generate income losses as well as capital ones through the use of swaps and other expenses.”) And:

The tax-aware strategies have helped AQR to add $47bn since March last year and Quantinno $39bn since January 2025, according to investment presentations seen by the FT and people familiar with the figures. That compares with growth in the wider hedge fund industry of about $628bn last year, according to data provider HFR.

Almost a third of AQR’s total assets were held in its tax-aware funds by the end of December.

One way to think about it is that, in a lot of the hedge fund industry (pod shops, etc.), what you are selling is “pure alpha” — outperformance not correlated to market returns — and there is only so much capacity to generate pure alpha. In the tax-aware industry, what you are selling is “credible alpha” — a reasonable amount of alpha, an IRS-placating quantity of alpha — plus the market return, plus tax losses. You can spread some alpha over a lot of money. So, more capacity.

Also:

Some investors and advisers also warned of potential regulatory scrutiny. “The deepest and most fundamental risk is that the IRS catches wind of this,” said one financial adviser.

Shh.

Elsewhere, Victor Haghani and James White at Elm Partners have a paper on “Robbing Peter to Pay Paul: A(nother) Look at Long/short Direct Index Tax-Loss Harvesting.” “Long/short Direct Index Tax-Loss Harvesting” is the schematic trade that I described here, the one you are not supposed to do (or admit to doing), the one where you are long the index return and short the index return, without any stock-picking alpha. They are not particularly bullish on it (because of fees). It’s better, both for your returns and for the IRS, if you try to pick mostly the good stocks.

Back in 2021, I wrote about a Bohemian prince who wanted to renovate his castle. He needed money. Traditionally there are two ways to get people to give you money to renovate your castle:

1. You could give them something valuable in exchange for their money. The castle was stuffed with Old Masters; sell some of those and use the money to gussy up the rest of the castle.
2. They might give you the money freely, out of the goodness of their heart, or to befriend a prince, or to associate themselves with an important cultural institution (the castle). There is a well-established economy of people donating money to cultural institutions, you’ve got a cultural institution, you can solicit donations.

But this was in 2021, a real high point for crypto, and the prince came up with an approach that combined the best of both worlds: He sold nonfungible crypto tokens, digital records on a blockchain recording, essentially, that people bought the tokens. I wrote:

Would you like to chip in a few hundred euros to help a 27-year-old Harvard-educated Bohemian prince restore some of his family’s castles? … Obviously he would keep the castles and paintings; after all he is a prince and you’re not. But you would get the warm satisfaction of helping out a prince in his time of relative need.

Is that not an appealing pitch? Well, okay, let’s add a sweetener. He will give you a receipt. If you keep a receipts collection, you can put this one in your receipts collection. You can show it off to people. “This receipt represents the time I gave 100 euros to a Bohemian prince to restore one of his castles.” Perhaps people will be impressed. Perhaps they will be so impressed that they want to buy the receipt from you. Perhaps they will pay you more than you paid. “Oh wow, I wish I had a receipt like that, tell you what I will pay you 200 euros for it.” ...

Perhaps a robust market will develop for these receipts. Who knows how much these receipts might be worth in 10 years? ... Early contributors — early investors in the prince-donation-receipts asset class — could become billionaires as their receipts appreciate in value. Honestly you can’t afford not to chip in some money, to help this prince restore his castles, and get a receipt.

There is something deeply odd about all of this. If you give a company money in exchange for a share of stock, the stock represents not just a historical record of you giving the company money, but also an ongoing economic claim on the company. If you sell the stock to someone else, that person gets the ongoing claim. It makes perfect sense for there to be a robust and volatile secondary market.

Whereas the prince’s NFTs represented no ongoing economic claim. You might buy the NFT from the prince for the same reason someone might donate money to an art museum: You want to be a patron of the arts, out of some combination of public-spiritedness and social climbing. But someone buying the NFT from you isn’t helping the prince restore his castle, and he’s not getting any ongoing claim on the castle. He is getting purely the historical record of your donation.

In 2021, this was a thing. In hindsight, it feels like maybe people were confused. They wanted to give money to the issuers of the NFTs, but they also thought they would get rich off their tokens. The purpose of the NFTs was a bit blurry; why couldn’t they be both a cultural donation and an appreciating asset?

Anyway. World Liberty Financial is a crypto project that, possibly among other things, provides a way for people to give money to Donald Trump and get a receipt for it on the blockchain. Why would you want to give Donald Trump money on the blockchain? Reasons! The goodness of your heart! One of the biggest buyers of World Liberty tokens, when they launched in 2024, was Justin Sun, a crypto entrepreneur who at the time was being sued by the US Securities and Exchange Commission for fraud. Now he isn’t! 

What did Sun get in exchange for buying $30 million of World Liberty tokens? Well, possibly among other things, he got the tokens. Which are tradable. Except lol they aren’t:

World Liberty Financial Inc., a Trump family crypto venture, is facing an investor revolt that includes billionaire backer Justin Sun, who accused the project of secretly building controls that let insiders freeze token holders’ funds.

Sun, who poured tens of millions of dollars into World Liberty, called the project “a trap masquerading as a door” in a post on X on Sunday. …

WLFI token’s value has dropped by more than half since a portion of its supply was unlocked for trading last year. It hit an all-time low below $0.08 over the weekend, according to CoinGecko data.

Again I feel like there was some confusion here.

Briefly the way shareholder voting works is:

1. It is absolutely irrational for a normal individual shareholder to pay attention to the governance of a public company, or to vote her shares — to elect directors, approve corporate actions, or vote on shareholder proposals — at the shareholder meeting. Her vote doesn’t matter to the result, or to her own wealth.
2. But the company has to decide what to do. How can it make decisions, when the shareholders — its owners — don’t pay attention or participate in governance? 
3. The answer is that the shareholders have to delegate the decision-making to someone else.
4. Traditionally shareholders delegated decisions to the company’s board of directors: The directors are nominally elected and supervised by the shareholders, but in practice they sort of appoint themselves and use their own judgment to run the company. At the shareholder meeting, the board tells the shareholders how to vote, and the shareholders rubber-stamp the board’s decisions.
5. A more recent approach is that shareholders delegate power to big asset managers: Instead of buying 100 shares of stock in a dozen companies, a retail shareholder will buy shares in a mutual fund or exchange-traded fund, and the fund’s manager will vote the shares and try to supervise the board of directors. Sometimes the asset managers will be annoyed with the board, and will vote against its recommendations — or even back proxy fights to replace the board — at the shareholder meeting.
6. There are reasons to think that the asset managers will do a better job of looking out for individual investors than the board will, and there are other reasons to think they will do a worse job.

When I first encountered these issues, in the 2000s, the asset managers were on the upswing: “Of course companies should answer to shareholders,” people would say, and by “shareholders” they meant professional asset managers. In recent years, the boards of directors have regained the upper hand, and now it is considered a bit shameful for big fund managers to try to tell companies what to do.

Maybe the starkest example of the resurgence of board power is at Exxon Mobil Corp., which last year implemented a program that allows retail shareholders to delegate their votes to the board. Instead of just throwing away their proxy statements, or trying to figure out how to vote, the shareholders can just check a box saying “whatever the board wants is fine forever,” and the board can vote their shares. This was controversial among the sorts of people who like good corporate governance, because “good corporate governance,” in the standard 2000s sense, means making the board answerable to robust shareholder power. But also it’s fine? Like, if you are an individual retail shareholder of Exxon, it’s probably because you trust the management of Exxon more than you trust, like, BlackRock Inc. So you will probably be happy letting the board vote your shares, to counteract the voting power of BlackRock. (And if not, you don’t have to; the program is optional.)

Still, as a matter of tidy-mindedness and comic irony, you might find this program annoying. “Why,” you might ask, “does Exxon offer retail shareholders the option of always voting with the board, and not offer them the option of always voting against the board?” Of course there are straightforward answers to that:

• Exxon wants you to vote with the board, not against it. 
• You probably want to vote with the board, not against it, if you are a retail Exxon shareholder.
• The board has fiduciary duties to shareholders and will probably take mostly reasonable positions, whereas people can submit all sorts of crazy proposals; if you always voted against the board’s recommendation, you’d sometimes vote for crazy stuff.

Still, tidy-mindedness and comic irony are powerful forces. Last week Exxon filed the proxy statement for its 2026 annual shareholder meeting, and it includes this shareholder proposal submitted by the New York City Comptroller’s Office:

Resolved: Shareholders request that the Board of Directors adopt and disclose policies to ensure that Exxon Mobil Corporation’s (‘Exxon’) retail shareholder voting program provides multiple independent options to shareholders so that the retail voting program does not inordinately advantage the Board’s own voting recommendations. The offerings to retail investors may include a range of possible voting options, such as: independent voting options based on standing instructions; a general ‘against management’ policy; and/or customized policies.

I feel like the board would say: No, the point is to inordinately advantage our own voting recommendations, because we are fiduciaries for shareholders and our recommendations are good. (It basically does say that: “Changing the Voluntary Retail Voting Program to compel the Board to include language or options in its solicitations that are against the Board’s recommendations, whether directly, through independent voting instructions, or a customized voting policy managed by a proxy advisor, is inconsistent with state law and the Board’s fiduciary duties.”) This is not actually a way to run a company. But as a matter of trolling, you have to respect it.

Elsewhere: “Shareholder proxy proposals have hit a five-year low as environmental and social resolutions experience a prolonged decline amid the US regulator’s efforts to limit investor activism.”

When you think about it, a lot of Tesla Inc. engineers were sent to work for free at Twitter Inc. when Elon Musk bought it, so in a sense Twitter owes Tesla a favor. Twitter, of course, is now called X, which is part of XAI, which is part of SpaceX, which is targeting a $2 trillion (?) initial public offering in June (?). So, transitively, SpaceX owes Tesla a favor. Anyway Bloomberg’s Dana Hull reports:

Sales of Tesla Inc.’s Cybertruck have been propped up in recent months by Elon Musk’s other companies, an unusual arrangement that further indicates the polarizing pickup is failing to appeal to everyday buyers.

SpaceX, the Musk-led rocket and satellite maker, accounted for 1,279 — or more than 18% — of the 7,071 Cybertrucks registered in the US during the fourth quarter, according to registration data that S&P Global Mobility provided to Bloomberg News. The billionaire’s other ventures acquired another 60 vehicles during those months. …

Photos and videos have circulated online showing long rows of idle Cybertrucks on SpaceX property in Texas. The lead engineer for the pickup posted on social media in October that SpaceX was replacing gas-powered support vehicles with trucks. At least some are being used as security vehicles. EV news outlet Electrek reported in December that SpaceX could ultimately buy about 2,000 Cybertrucks.

Also won’t they look cool on Mars, when SpaceX colonizes Mars?

At this point, there is not much more to be said in the vein of “boy, Elon Musk shifts resources among his companies without much regard for traditional corporate niceties.” But he does!

Sure?

Allbirds Inc.’s surprising pivot from wool sneakers to artificial intelligence infrastructure is looking more like a flash in the pan for the stock than a driver of long-term gains.

Shares of the firm, which said it plans to re-brand itself as NewBird AI, sank as much as 31% on Thursday, a swift but only partial reversal from the knee-jerk reaction a day prior when the stock soared more than 582%. Nearly 300 million shares exchanged hands on Wednesday, several times the daily average of just over 20 million.

“This has the feel of a meme stock, where emotions take over and logic and reason get thrown out the window,” said Adam Sarhan, chief executive of 50 Park Investments. “That the market actually rewarded the stock yesterday when it doesn’t seem to have any kind of actual AI edge tells me that froth, specifically AI froth, is picking up.”

I dunno. At noon today NewBird was trading at about $11.92 per share, which is still up more than 375% from where it was on Tuesday as a defunct sneaker company. Because yesterday it announced that it got an investor commitment of five million dollars [4]  to build AI infrastructure. Sam Altman is out there committing trillions of dollars for data centers, and Allbirds is like “we have $5 million??” and the stock moons. Sure!

Anyway the main lesson that I drew from this yesterday is that more micro-cap public companies should try this. That is obviously a gloomy lesson, but cynical AI pivots get rewarded and presumably someone is making money off of this. (Probably the investor who committed the $5 million.) On that note:

Myseum, Inc. (Nasdaq: MYSE) (“Myseum” or the “Company”), a privacy-first social media and technology innovator, [yesterday] announced that it is now operating under the new name Myseum.AI, Inc. effective April 15, 2026. The rebrand illuminates the Company’s core technology platform that will integrate proprietary privacy-first artificial intelligence (AI) into its secure messaging and social media platforms. The Company is developing privacy-first agentic localized AI agents that assist in managing personal media such as photos, videos and messages, while maintaining privacy. 

At noon today, Myseum was up 125% over yesterday’s close. Sure!

US Probes Suspicious Oil Trades Made Before Trump Pivots. Blue Owl Surges Most Since 2022 Amid Banks’ Private Credit Calm. “Losses from [Goldman’s] US nonlinear gamma rates desk ...were one of several factors that contributed to the division missing analyst expectations by nearly $1 billion in the first quarter.” Spain’s Repsol wins back control of Venezuelan oil operations. SpaceX Plans Site Visits for Large Investors as Mega-IPO Nears. How Anthropic Learned Mythos Was Too Dangerous for the Wild. Live Nation Illegally Monopolized Concerts and Ticketing, Jury Finds. EU to relax merger rules in bid to create ‘European champions.’ Wall Street Quants See an Edge in Polymarket Earnings Forecasts. Barneys might come back.

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!