The Politics of Box Spreads
Why some people on “Main Street” must start thinking about money markets the way everyone does on “Wall Street”
If you’re a finance person, you think a lot about the rules of finance, which will appear to you very much like what Marx called “laws of motion” – because they structure the markets within which you operate on a daily basis. But your main concern is not really how the rules got made, and it is only tangentially how they might get altered; you focus primarily on how to use the rules as leverage to create trading profits.1
Just as Foucault famously said “where there is power there is resistance,” so every trader knows that where there is regulation there is the opportunity for regulatory arbitrage. If the trader can get around the intent of the regulation without actually breaking the law, then they can make money (for their firm, for their client, perhaps for themselves).
If you live in, or visit, the world of finance, you at least saw and probably read about a big story a few weeks ago. Outside of that world, this big story was a non-story; you almost certainly heard nothing about it. Bloomberg broke the news, announcing the launch of a new exchange-traded fund (ETF) that implements a trading strategy called “box spreads” in order to produce “guaranteed” returns that do not count as interest income under US tax law. Compared to news outlets like NY Times, which frequently refuse to talk about capitalism, even when they are talking about capitalism, I find it refreshing to read the financial press (the FT and Bloomberg, especially), because they are often so much more honest and direct. Here’s the lede in Bloomberg’s box spread etf article: “A Marine Corps veteran with a finance Ph.D. has come up with a new way to avoid taxes.”
That really is the heart of this story, but I want to partially unpack it, because it offers a recent – in some sense banal, yet also acute and significant – example of the politics of money markets. In other words, it exemplifies what this newsletter is all about, i.e., money/power.
1) Money and risk
Before turning to box spreads themselves, we have to start with the fact that money, as the claim of a creditor on a debtor2 can never be guaranteed; it always remains precarious and uncertain. Money is a claim on value in an unpredictable future and is therefore always risky. Any example of a money-credit (from a $20 bill to a soybean futures contract3) combines numerous forms of risk. Modern finance has a compendium of ways to separate (or appear to separate) types of risk, to build new forms of money (financial-asset-products) that attempt to concentrate one form of risk and eliminate others. In particular, and most fundamentally, some simple financial engineering can be used to hive off “credit risk” from “interest-rate risk.”
What’s the difference between the two?
Credit risk is easy. If I borrow $80 from you for a year, at an agreed interest rate of 3%, you now have a money claim against me. The loan contract (or merely the note I sign that says, I owe you $82.40, payable on 1 March 2025, ––Sam) is itself the money (what Keynes called the “money thing”) – the token that represents your claim on me as your debtor. Your money, like all money always and everywhere, is therefore at risk. First, and most obviously, because, for a variety of reasons, I might not pay you back. I might shut down my Substack tomorrow and move to an unnamed island to live off the money from your loan to me (in the form of your paid subscription to “Money/Power”). This is credit risk, and it’s what most of us think about first, foremost, and perhaps exclusively when we hear the phrase “risky loans.” This typical intuition led, for example, many commentators to totally mischaracterize the collapse of Silicon Valley Bank (SVB), blaming the fat-cat bankers for making too many “risky loans” that were bound to default.
Of course, that’s not what happened at all. SVB’s collapse was a story of interest-rate risk, which is much less intuitive. To grasp credit risk you only have to know what it’s like not to get paid back when someone owes you. But we can only grasp interest-rate risk if we think the way Wall Street thinks – the way money market participants and the world of finance thinks. In that world, every single financial asset (what I tend to call money-credits – claims on debtors of some sort) can be written down on a spreadsheet with a present value.
We can get the present value of a commodity simply by checking its current market price. But the present value of a financial asset, of a money claim, looks quite different, because money itself has no intrinsic value in the present – its value is always a future claim on value. You know the future value of the loan you made to me ($82.40), and according to the loan terms, you won’t actually receive the money for a year, so it seems as if it has no present value. However, both the logic and practice of finance and money markets are relentlessly driven to calculate a present value for every money claim, as a first requirement for trading that claim. In other words, what if a month after loaning me the money, you decide you need cash, and thus you want to sell my loan to a friend? How do you know what $83.20 in 11 months is worth today?
The answer depends on the current interest rate. You loaned me money a month ago at 3%; tacitly, but quite powerfully, you were saying that your $80 today was worth $82.40 in 12 months (and that $82.40 in 12 months is worth $80 today). But what if interest rates surged over the past month (say, because the Fed was raising rates), such that today your friend can buy a CD or T-Bill and earn 5% interest? She thinks of her $80 today as being worth $84 in 12 months. She’s therefore not going to pay you $80 for something that’s only going to be worth $82.40 a year from now.4 In other words, in deciding the present value of the loan you want to sell her, your friend does not directly care about the 3% interest rate you and I agreed upon. She cares about the current market interest rate (5%) and the future value of the loan ($82.40). She will calculate the present value of your loan by figuring out how much money she would need to loan to someone today, in order to get back $82.40 a year from now. The answer is $78.47.
I have been offering a non-standard – some might say weird – explanation of two key concepts in finance: present value (in this case, $78.47) and the discount rate (here, 5%). If you do an internet search5 on these terms you will get much shorter answers including a simple formula for PV, but I spent the time to write the above words (and ask your forbearance in reading them) because I think those short answers are shockingly unhelpful. In my experience, search-result answers tend to convey a kind of transhistorical truth about the concepts, as if the present value of a financial asset was somehow intrinsic to the asset. It’s not. Present value and the discount rate only have meaning within the context of advanced capitalist money markets, wherein every money-credit is tradable because someone makes a market in that type of money-credit.6 In our example, you only care about any of this because you are trying to sell the loan. Your friend does the present-value calculation not because she herself is directly interested in loaning you money (as you, I imagine, were when you loaned me the original $80), but because she wants to maximize yield. She’s thinking like an investor.
The example draws into stark relief the palpable truth of interest-rate risk: you loaned me $80, and a month later the loan itself is only worth $78.47. Your money-credit (the loan, which is a claim on me as your debtor) is worth less because interest rates went up. The failure of SVB occurred primarily because of this exact logic: they made loans that were not at all risky in terms of credit, but turned out to be very risky indeed in terms of interest rates. SVB held $91 billion in low interest-rate bonds on its books, but when interest rates rose the present value of those bonds (or in redundant wall-street speak, the “mark to market value”) fell to $76 billion. The details of SVB’s actual collapse are complicated, but the underlying logic is simple: that $15 billion loss made SVB technically insolvent (its liabilities exceeded its assets) and when their depositors caught wind of this, they started a run on the bank – leading to the most massive withdrawal of funds ever ($42 billion in one day alone) – and rendering SVB illiquid.
Interest-rate risk is real because money is never stable or guaranteed. Stable money is a contradiction in terms. To hold money is always to be subject to credit risk, interest-rate risk, and many other risks besides.
2) “Interest” and “capital gains” in the US tax code
Many readers may now be wondering why they, or anyone not in finance, should care so much about the nature of interest-rate risk and its difference from credit risk. My answer: because US tax law makes a profound (arguably invidious) distinction between the two. The tax code first assumes it’s possible to cleanly separate the two risks, and then it punishes people who take on interest-rate risk and rewards those who take on credit risk. We can mostly ignore why the law developed as it did, except to say that I suspect it has something to do with the fact that those who lobbied for these laws often based their arguments on a false concept of money – one that reserved a special place for sound money, which it distinguished from the bold and courageous, risky investments of founders and entrepreneurs.
The idea is that if you put your money in a savings account or CD, buy a sovereign or corporate bond, or even make a loan to a friend, then you are just earning interest, whereas if you buy stock you are taking risk. The IRS taxes (that which it defines as) interest income as “ordinary income,” which means you pay your marginal tax rate (up to 37%), but it then creates a new category named “capital gains” – the appreciation of assets – subject to a distinct capital gains tax rate (up to 20%).
Here’s Matt Levine on the difference:
Interest is what you get paid for giving up the present use of your money. If you put your money in a savings account, the bank pays you interest every month for the use of your money. But if you buy, like, Nvidia Corp. stock, or Bitcoin, nobody pays you any interest. Instead, you expect the value of those things to go up over time, to compensate you for the use of your money.
Levine is always lucid, but notice that his account here seems confused. He is attempting to draw out the tax law’s sharp distinction between “interest” and “capital gains” but in both cases he refers to compensation for forgoing “the use of your money.” The confusion here isn’t Levine’s; it’s baked into the IRS code itself. Buying a bond or stock both involve giving up your money now in hopes of gaining more in the future. The difference between “interest income” and “capital gains” cannot be found in the nature of different money-credits; they are constructions of the tax law.
Constructions are never not political, though they are also never strictly political. Hence the practical and historical justification for the distinction involves both economic and political forces/relations. For our purposes what matters is that the argument advocating a separate (lower) capital gains rate hinges on the difference between interest-rate risk and credit risk. The tax law is designed to encourage and reward the latter.
Often tacitly assumed, and sometimes explicitly stated, the core idea is that only credit risk is real risk. Thus, there must be a “risk premium” to encourage investors to put their money in stocks, commodities, and derivatives. The risk premium takes the form of a lower tax rate, and the tax savings are explicitly designed to serve as “compensation to you for taking the risk that [the stock] might go down instead” (Levine, 22 Feb).
There’s the rub.
Everything might go down. In capitalist money markets, the present value of financial assets changes by the day, the minute, the second. There is no such thing as a money-credit that cannot go down, because under capitalism, interest-rate risk is just as real as credit risk, foreign-exchange risk, etc. Everything is risky.
The notion that we must incentivize investors to make “risky” investments proves false on both sides: on the obverse, because investors can never hold “non-risky” money-credits; on the reverse, because there is nothing inherently “better” (for society) about buying a stock rather than a bond. Here I respond to the idea that a lower capital gains rate encourages investment in real economic activity. This oft-repeated idea, which serves as the strongest political argument for a lower capital gains rate, has been repeatedly empirically debunked, but the argument is itself logically untenable on its face. It should be obvious that whether I buy shares of twitter stock or a tranche of twitter bonds, I’m still giving Elon Musk my money, and he can do with it what he pleases. Bond purchases fund real economic activity just as much as stock purchases do.7
But this argument is profoundly wrong in much more interesting and important ways. The intuitive idea of a lower capital gains rate is that I should not be rewarded for putting my spare cash in a savings account, but that I should be rewarded for starting a company or buying stock in one. The former produces idle interest income; the latter is dynamic and innovative investment. The tax code insists we hold the line between safe deposits and risky investments.8
In this way the tax code thoroughly misunderstands and ignores the raison d’être of banks: “The bank is a machine for turning safe deposits into risky investments.”9 Banks do not hold my money in trust: my bank deposits are my loans to the bank. The bank borrows from me in order to lend to others. And in order to make money doing so (and a bank today must make money doing so, otherwise it collapses), the bank will take on more risk. My “safe” deposits are funding for the bank’s “risky” loans.
The conclusion here is a mirror image of my conclusion to the first section: once we grasp the nature of money as an always risky claim on future value, we see that the tax code’s unequal treatment of some money returns compared to others is a political construction, untenable on strictly “economic” grounds (and just plain wrong on political grounds).
3) What has any of this got to do with “box spreads”?
We have seen that in any concrete case of money market funding for economic activity, there will always be a variety of risks (and never a zone of safety). But contemporary finance wizardry makes it possible to create crazy synthetic money-credits that maximize the separation between credit risk and interest-rate risk (as well as other forms). In other words, through the magic of derivatives, today’s investor can choose to maximize or minimize one form of risk over another.
To see how, we need to start with the fact that even the simple act of buying and holding a stock involves interest-rate risk, because holding a stock for five years “ties up my money” in the same way as buying a five-year CD from my bank. Therefore, despite IRS treatment, if I buy a stock today and sell it in 2029 for a higher price, part of my profits will be a form of “interest” on my money (though the IRS will treat all the profit as capital gains). Money market traders understand this, and they can do what the IRS cannot: separate the interest-rate risk from the credit risk.
This means that rather than loan my money to a bank (deposits) or a company (bonds) I can effectively loan it to the money markets (liquidity). I do this through derivatives, by entering simultaneously into four different options contracts, all on the same stock, and all with the same expiration date: I both buy and sell a call option (the right to buy the stock) and I both buy and sell a put option (the right to sell the stock). These four contracts are the four corners of “the box.” Interested readers can find the relevant details here, here, and especially here,10 – and can see the basic structure in Figure 1, below. However, for the purpose of this discussion, all that really matters is that at the expiry date, no matter what the price of the stock is, two of those contracts will be exercised (while two will simply expire) and the result will be a return to me that approximates the current “risk-free” interest rate (e.g., around 5% at this moment in time).
I am, for all intents and purposes, making a loan and receiving a return (interest) just as much guaranteed as it would be if I bought US T-bills.11 But the whole point of this financial engineering is to create an interest-bearing loan that doesn’t look like a loan. I took on interest-rate risk almost exclusively, but because I bought and sold derivates, I will file my taxes by claiming capital gains, not interest income.
More accurately, “I” as individual investor won’t do any of this. Rather, the box spread strategy will be implemented by Alpha Architects, and all of this will happen inside the ETF. I will merely buy shares of the ETF. As Levine shows, box spreads are a great strategy, but “not a great product.” Moreover, while the options strategy is designed to turn interest income into capital gains, the ETF is designed to avoid capital gains taxes.12
Let me be clear: box spread strategy is kind of brilliant. Only some incredibly smart people could have come up with it; it took a lot of thought, effort, and creativity, and the end result has a certain beauty to it. Levine gives it his highest words of praise – “pretty good!” – and in one sense I cannot really disagree. But at the end of the day, what does the box spread really do? It creates a distinct new money-credit that has only interest-rate risk but can putatively13 be categorized for US tax purposes as the opposite (only credit risk). And what are its concrete implications?
In a word, one that has become something of an epithet in American politics, redistribution. The money that goes into the Box ETF is money that would have been going into T-bills, money-market funds (which invest in T-bills), CDs, and high-yield savings account. It will have no discernible effect on individual investor decisions, and negligible impact on the allocation of capital. The only real result is that some money that would go to the US treasury for payment of taxes on interest income, will instead get divided up between the ETF managers and individual investors.
Boiled down to its desiccate, regulatory arbitrage always amounts to profit for traders and loss of tax revenue for the government.
The regulators will always be disadvantaged in relation to the traders, because the latter can earn billions of dollars in profit if they “win,” while the former will still just get their government salary. The regulators therefore need more democratic citizens on their side.14
Figure 1: Where’s the Politics?
The CEO of your hedge fund, the President of your investment bank, and a host of other “higher-ups” may think about this a lot; in fact, they may devote much of their time and money to getting the laws changed in ways favorable to your firm. But as a trader or agent of a money market dealer, your main concern is not setting/changing the rules but “working within them” or “gaming the system” (these phrases have radically different connotations and very similar denotations).
Money can be better grasped as a worldly phenomenon if we think about it not as a thing (it is not a thing) or any kind of simple relation (it is a relation, but not a simple one), but as an array. The array includes 4 elements: creditor, debtor, the money token itself (something that represents or symbolizes the relation between creditor and debtor), and finally, denomination. The credit/debt relation must be both named (euros, pesos, yen) and counted (€47, ₱4,700, ¥47,000,000).
Yes, I argue that tradable derivatives contracts are absolutely forms of money. For readers that this strikes as insane, please know that I will return to this topic in future posts.
Roughly a year from now. In order to make this example a helpful and illuminating heuristic, I’m ignoring the difference between the 11-month and 12-month loan durations.
I think we have to stop using google as a verb, because, as everyone now knows, google search results have become so enshittified that many people, myself included, have stopped using google for search. Many people are shifting their search needs to ChatGPT or other LLMs. (I started using Kagi a few months ago, and it has been a revelation.)
Not to mention the fact that there is never one interest rate (singular) and the multiple interest rates (plural) are always determined by economic, political, social, and cultural forces.
Bankruptcy law dictates that should a company go bust, bond holders get paid back before stock holders. Does that make bonds a safer investment? Sure. Although context is everything, because junk bonds are arguably much riskier than blue-chip stocks. In any case, even focusing on one company, the investment in bonds still provides funding, so there’s no political reason to privilege the stockholder. Moreover, while the bondholder’s downside might be somewhat limited, so is his upside (he’s only ever going to get his money back, plus interest). In other words, there’s already an economic incentive in place to favor stocks over bonds: more risk, more reward. Capital gains tax rates can only amplify, not create, an already extant incentive.
Unsurprisingly, the tax code itself constantly blurs that line: if I buy a bond and hold it to maturity, I’ll receive regular coupon payments that the tax code will count as interest. If I buy a bond and then interest rates move in my favor (i.e., down) I can sell it for a profit, which the tax code will then treat as a capital gain (appreciation of an asset).
Much more on this in the promised future post, “What is a Bank?”
Levine himself (the third source) draws from the first two extensively, and I would point to Levine’s second numbered list – starting with, “1. You don’t buy any stock” – as the most clear and concise explanation of the options details. Note, however, that the Alpha Architect piece is written by Wesley Gray, the CEO of the new ETF, and the Morningstar article cites as its main authorities both the Alpha Architect piece and Gray himself. All of this is really coming from one guy, who has an enormous financial stake in getting more people to put money in his ETF. Again, shout out to Bloomberg for heading their story with a picture of Gray and the caption, “Wesley Gray in Huacao, Puerto Rico.”
My counterparty is therefore taking on all the credit risk; they get all the upside or downside in the stock movement, and their long exposure is leveraged, because they did not buy the stock up front (they effectively borrowed the money from me, at 5% interest, in order to buy the stock synthetically). Crucially, my counterparty in the options market is not another individual trader: it’s the Options Clearing Corp. (OCC), a major market maker. While the OCC is not directly backed by the government, it is, as Gray himself says “a SIFMU, or systematically important financial market utility.” In other words, none of this is possible without what Martijn Konings calls “the bailout state.” And, as always, “riskless assets” are never riskless, but rather subject only to the risk of governmental default (or in the case of US Treasuries, global capitalist crisis).
Not entirely, of course: when I sell the ETF I may realize gains, but the ETF will be buying and selling money-credits all along the way. It will use magic to make that buying and selling look like something else (namely, creation and redemption of ETF shares).
Some have argued that Section 1258 of the Internal Revenue Code could be used to argue that the BOXX ETF is what it is – interest income, not capital gains.
No one should read “Money/Power” to find investment advice; I do not and will not provide any. Still, do not put your money in boxx spreads.