The Box Spread Trick: Get rich slightly faster
Disclaimer: This is not financial advice. Also, following these steps slightly wrong could result in you losing 100% of your savings. I have listed what I think are the most likely pitfalls, but still, beware.
Thanks to gilch and Mark Xu for helping edit this post, and Wei Dai for the original idea.
If you have a non-IRA brokerage account, you can borrow money on the open market at 0.55% and put it in a certificate of deposit (CD) yielding at least 1.25%. After taxes and commissions, this increases annual returns on your entire portfolio by 0.2-0.5%, nearly risk-free. This opportunity could continue for several years.
As this is $1000-2500 a year on a $500,000 portfolio, it’s probably worth your time to set up if you’re mid- to late-career, especially if you already do things like credit card churning. The administrative work includes making one trade every 1-3 years and a small amount of extra tax work.
Before actually performing the trick, there are a few things you need:
A taxable (not IRA/401k) investment portfolio worth more than about $100k.
A brokerage account with portfolio margin and options approval. Portfolio margin is not strictly necessary but increases the return. I suggest using Interactive Brokers or TD Ameritrade.
An account at an FDIC-insured bank (or NCUA-insured credit union) with a high-yield CD or savings account. You can find a list of such banks at depositaccounts.com, and approval is fairly quick. FDIC/NCUA insurance is limited to $250,000 per individual per bank, so if your portolio is over $500k you might need multiple banks.
In total, the setup took me about 50 hours, but this involved lots of trial and error and double-checking; it can be done in 10 hours.
The Box Spread Trick
As of September 2020, some banks offer a 3-year CD yield of 1.25%, while the Treasury yield is only 0.2%. If you could borrow at the same rate as the US Treasury, you could make 1.05% per year risk-free. Sadly, you can’t do this conventionally: mortgage rates are upwards of 2%, and most brokers charge upwards of 3% on margin loans. But you should be able to, because you have collateral in the form of a stock portfolio. This is where the box spread comes in.
A box spread is a combination of four options which cancel each other out so there is no risk from market movements. This allows institutions to lend each other large sums of money: one party sells the box spread for a premium, and pays the loan back on the exercise date up to 3 years later. As a retail investor, you can’t withdraw the cash from this loan, but you can take out a margin loan from your broker which is financed by this box spread. By using box spread financing, you pay market rates (typically treasury yields plus ~0.3%) rather than the broker’s 3-10% margin rates.
To perform the trick, sell SPX box spreads in a total amount that’s 30-65% of the value of your account depending on how frequently you want to monitor it. The exercise date should be as far in the future as possible, currently 2-3 years. Each box spread is worth 100x the spread width; i.e. $10,000 for a 2900⁄3000 box. The legs should be 100-300 points apart to minimize commissions cost, and the upper strike price should be near-the-money (near the current value of the SPX) to maximize liquidity. In current market conditions, the trade looks like this for a $250,000 portfolio; it’s a loan with $150,000 due in December 2023.
SELL 5 DEC 15 2023 3000/3300 SPX BOX
WARNING: Don’t trust a spreadsheet someone on the internet made. Double-check the math yourself. Also, make sure that:
You’re using a LIMIT order (not market). Bid-ask spreads are absurdly wide.
You typed in the price correctly.
You’re selling European-style options, e.g. SPX. This prevents early exercise risk.
You’re selling the box, not buying it. (An order to sell a 3000⁄3300 box is equivalent to buying a 3300⁄3000 box; this is fine. The preview price should be negative, or say CREDIT).
You’re doing a single trade, not four separate trades.
When previewing the order, the profit graph is a perfectly flat horizontal line, indicating that the options cancel out.
Execute the trade, and wait for it to be filled. If it isn’t filled within a day or two, slowly walk up the price until it is. (I was filled at 0.32% above the treasury yield.) You may get a slightly better price or fill time by direct-routing to an exchange rather than using your broker’s “smart routine”. Once the order is filled, deposit the money into your bank account/CD. Your brokerage account should now have a cash balance near zero.
If the value of your investments declines enough during the 3-year period, you could be subject to a margin call and have your investments liquidated. Before this happens, withdraw some money from the CD (paying a small penalty, typically the last six months’ interest), and redeposit it into your brokerage account.
Let’s say you have $350,000 in a Vanguard brokerage account, which is invested in mostly broad-based indices with some narrow-based indices, say mostly VT with some VWOB. You get an Interactive Brokers account with options trading and portfolio margin, and pay a $50 fee to transfer your assets.
SELL 10 DEC 15 2023 3200/3400 SPX BOX
Say the Treasury yield on September 16, 2020 is 0.17%; you target a rate of 0.46%, and calculate the price. You make the trade above for $197,000 (0.471% APY); after a $26 commission, you get a $196,974 credit. The loan is due in 3 years for $200,000. You have already set up an account in a credit union with a 1.25% Jumbo 3 Year CD rate, so you withdraw the $196,974 from IB and use it to create a CD maturing in December 2023.
The liquidation value of your IB account is now $150,000 and your exposure is $350,000, a 43% ratio, comfortably above the 15% requirement. If the market declines by 28%, you could be subject to liquidation, so you set up a phone alert on the IB app.
In December 2023, the options are about to expire, so you buy them back or let them be exercised for $200,000. By then the CD has accumulated $8,164 in interest, which is $6,182 after-tax (your tax bracket is 24%) while the box spread has lost $3,026; this is $3,134 in profit. You then sell another box spread expiring in 2026, repeating the process.
Is this really risk-free?
I’m moderately confident that the only major risks are (a) doing the setup wrong, and (b) having your options partially liquidated if the market plummets by 30%+ before you notice. Having part of a box spread liquidated is bad for two reasons: first, bid-ask spreads are wide, so you’ll be liquidated at unfavorable prices, and second, holding part of a box spread is extremely risky, often equivalent to >20x leverage. You can limit this risk by limiting the size of the loan or using Interactive Brokers’ “liquidate last” feature. You can eliminate it entirely by buying put options, though these are insurance and cut into your profit.
Why does such an arbitrage opportunity exist?
In short, people who put their money in savings accounts or CDs are less likely to withdraw it. Banks therefore give you higher rates. Wei Dai, who posted a comment outlining the trick in April, wrote that this effectively amounts to an FDIC subsidy. It’s not quite free though, since you are giving up liquidity.
CD rates are currently 1% higher than treasury yields. This is higher than historical averages, but the gap previously stayed around 1% between 2009 and 2013. Because FDIC insurance on CDs is limited to $250,000 per bank per individual, a 1% gap probably can’t be exploited at scale by institutions. The excess return could continue until market forces cause the gap to shrink, or until a large percentage of the money in high-yield CDs and savings accounts is from people using the box spread trick.
What if I pick stocks/own crypto/use Wealthfront/don’t have portfolio margin?
The portfolio margin requirement for narrow indices and individual stocks is 15% rather than 10%. If you don’t have portfolio margin, the Reg-T margin requirement is 25%. For cryptocurrency, no margin whatsoever is allowed. Actual margin requirements are often higher in volatile market conditions; sometimes brokers also have higher “house” margin requirements. Use your broker’s tool to adjust the loan amount according to your margin requirement.
Passive services like Wealthfront don’t support trading options, and so don’t support the box spread trick.
I’m not a tax professional, but I think LEAPS (the long-term index options we’re using) are Section 1256 contracts taxed as 60% long-term and 40% short-term capital losses. Interest from a savings account or CD is taxed as ordinary income, at up to 37%. The value of the box spread trick should decrease in higher tax brackets.
What if I’m outside the US?
The US has one of the strongest deposit insurance systems in the world. I haven’t seen non-US banks with similarly strong insurance and high CD/savings yields.
Interactive Brokers gives you options and PM approval on any account above $125k by checking some boxes. Rather than give a margin call, they immediately liquidate on a margin violation, but they offer a “liquidate last” feature so your options are not liquidated. They also offer lower margin rates, but this is not impactful because you don’t use the margin loan. However, their software is outdated and less user-friendly. TD Ameritrade has better software and (as of 2020) also has a bonus for transferring your assets there, but requires that you pass a written test on options. Other brokers might work (leave a comment if they do), but I’ve heard that Robinhood doesn’t let you trade box spreads. ↩︎
NCUA and FDIC are both backed by the full faith and credit of the US government, so the risk of bank failures is small. If you have a spouse, you can get one account for each of you, plus a $500,000 joint account; this lets you have $1M of total FDIC/NCUA insurance at just one bank. ↩︎
December options start trading in the September 3 years prior. ↩︎
I have heard that institutions use a 1000/2000 box, but these can only be exchanged in increments of $100,000 and so could be a bit unwieldy. ↩︎
This is a reasonable asset allocation if you mostly believe in EMH and want to approximate the global market portfolio but already own real estate and think Treasury yields are unreasonably low. ↩︎