Those who read my articles know that my investment philosophy revolves around trading against the constraints and biases of other investors. If everyone wants something shiny, I'm happy to sell it to them and buy underpriced assets to profit off their bias. I didn't think about it a ton when I read it the first time, but reading Michael Lewis's Moneyball changed my philosophy on investing.
I started studying the NFL and NBA and figured out that the winning teams tended to be those who used a more data-driven approach and had better discipline. Until I really dug into it, I had no idea the extent that investors beat themselves with their own biases. And if a lot of investors are going to insist on beating themselves with silly mistakes, inattention to detail, and a blind eye towards risk management, I might as well find ways to trade against them.
The trade I'm about to share is related to a constraint that a lot of large institutional investors have, which is their indifference to interest rate risk and an aversion to leverage that causes them to all buy the same overpriced assets.
Large institutions like insurance companies and pension funds like to match the duration of their liabilities to the duration of their assets. This is a sensible (but inefficient) strategy for investing in US Treasuries, which guarantees the investor that they will have X amount of money at X date, thereby guaranteeing the company that their liability will be satisfied.
Insurance companies and pension funds don't necessarily care how volatile the asset is in the meantime, because as long as they know what the payoff is in 20+ years, they're sleeping easy. This actually makes them the perfect kind of investor for us to trade against, because they're a huge part of the market that is indifferent to how risk is priced. You can make this trade yourself. It's likely to persist, as it's a structural flaw in asset pricing caused by economic incentives and not a price pattern that can be arbitraged away.
Institutional Investors Are Starved For Yield
Institutional investors pile into the same assets due to their inherent biases and constraints. Long-term bonds are probably the best example of this.
On the following graph, you can see the returns and risk for various Treasury maturities, as represented by their iShares ETFs.
SHY = 1-3 year Treasuries
IEI = 3-7 year Treasuries
IEF = 7-10 year Treasuries
TLT = 20-30 year Treasuries
As you can see from the graph, there's a normal relationship between risk and return for US Treasury ETFs, with a slight catch. Since US Treasuries are considered credit risk-free, the only risk priced in them is interest rate risk. The United States Treasury literally owes you the money they say that they will pay you over the course of the bond, so any deviation in the price of the bond is bound to revert to par value over time. Buying long-term Treasuries is a way to eke out some extra nominal return and still have it be guaranteed.
However, when you look at the risk-adjusted return of the ETFs, shorter-term Treasuries crush long term Treasuries.
The Sharpe Ratios are as follows:
SHY - 0.77
IEI - 0.75
IEF - 0.62
TLT - 0.45
As you can see here, investors are choosing to buy long-term debt over short-term debt in spite of the inferior risk-adjusted returns. And just in case you thought that I cherry-picked the time period, academic research shows that this risk/return disparity in US Treasuries has held since at least 1959. Interest rates have gone up, down, and sideways since then, but the anomaly has not even come close to going away.
Here's another white paper looking at the risk and reward of US Treasuries and how to exploit them from a cool niche financial advisory firm in Denver, Colorado.
Why Does This Anomaly Exist?
This anomaly likely has been allowed to persist because investors are so constrained from using leverage that they are actually paradoxically taking more risk by buying long-duration and overpriced assets than if they were allowed to use a little leverage.
Risk-adjusted returns go pretty much straight down as you go out on the yield curve. The study I mentioned earlier showed that T-bills seem to have Sharpe ratios close to 1 and require about 20x leverage to have the same risk profile as equities. 2-year Treasuries require about 7x leverage and have a Sharpe ratio of about 0.8. 5-year Treasuries have a Sharpe of about 0.7 in most cases and need 3x leverage. 7-10 year Treasuries only need 2x leverage but have a lower Sharpe ratio still. Finally, 20-30 year Treasuries have about the same Sharpe ratio as the stock market, as well as about the same standard deviation.
You could make the argument that 20x leverage on T-bills is too scary for someone who isn't a trader at an investment bank to make, and I'd agree with you. But a 3.75x position in 5-year Treasuries has the same risk profile as TLT with a much higher return, and since you're dealing with Treasuries, there's only so far that they can travel from the original price paid, since they're intrinsically linked to a future cash payment. I'm not aware of any idiosyncratic risk in the 3-7 year Treasury market that would cause this position to lose money.
As you can see here, 3.75x leveraged IEI beats TLT by almost double over the last 10 years with less risk. Because I set their durations equal to each other, both positions will lose approximately the same amount if interest rates rise. However, since the leveraged 5-year position is closer to the present, it makes more money from "rolling down the yield curve," which is simply the profit from the passage of time reducing the risk profile of the bonds. Here are primers on bond roll-down and bond duration if you need.
The bottom line: a leveraged position in IEI wins on every measure of risk and reward versus TLT. Leveraged IEI is in blue for this and all additional graphs.
Moreover, the outperformance of leveraged positions in short-term bonds has been shown since 1959, when interest rates were about where they are right now before going to the moon and back in the 1970s and 1980s.
I did additional out-of-sample tests that show that when interest rates rise, leveraged 5-year Treasuries have about the same downside as TLT, but when interest rates fall (when the Fed cuts rates), you make money faster and easier. But here's how the strategy did from 2016 to now, when interest rates rose.
Heads you win, tails you tie!
Here's how the strategy did the last time the Fed cut rates.
Again, leveraged IEI has less risk and more reward, including a staggering 43 percent return in 2008. Putting on a trade like this to complement your stock portfolio could have atoned for a lot of sins from August 2007 to the end of the crisis.
Note that the futures market won't exactly track the graphs I'm giving you. That said, the financing spread for Treasuries is roughly the same as the fees that the ETFs charge plus the risk-free rate. If I get some quality futures data, I'll happily share it, but traditionally for markets like Treasuries, there is little to no slippage.
How Do You Make Money From This Trade?
This trade needs to be made through the futures market (or the Treasury repurchase market if you have major money to play with). You could buy IEI on margin at Interactive Brokers, but the financing cost would eat up over 100 percent of your profit. So don't do that.
What you can do is go long 5-year Treasury futures 3.75x the amount you would otherwise buy in TLT. As short-duration US Treasuries are the best collateral lenders can ask for, the financing cost is among the lowest in the financial system. This solves the problem of paying a dime to make a nickel.
Also note that there are good reasons to have some kind of leverage constraint; for example, I'm not comfortable leveraging Treasury bills at 20x. I like 5-year Treasuries because there's a very liquid futures contract that you can trade on them, as well as the 2-year contracts (if you're a little less leverage-averse and want to profit from the current yield curve inversion).
The minimum contract size for futures is $100,000 for the 5-year contracts and $200,000 for the 2-year contracts. Your brokerage account should be about this size to be able to handle these trades as well as own stocks/ETFs.
There is additionally a massive tax break for trading bond futures. 60 percent of your profits from trading Treasury futures will be taxed as long-term capital gains, and 40 percent will be taxed as short-term. On average, this saves you about 15 cents for every dollar you earn from trading, because Uncle Sam allows you to. TLT, on the other hand, is taxed at the highest possible rate. Additionally, you can carry back losses 3 years and get a large tax refund if you have a losing year. If you've never had a capital loss carryover before, trust me, it's a huge perk to have this carryback in addition to the regular carryforward.
There are some times and places to use TLT, however. Your retirement accounts are a good place to have some TLT because you don't have to pay tax there, and simplicity and retirement go hand in hand. If you're comfortable with it, you typically will be able to trade futures in your IRA also. The risk/reward is obviously better, but you need to have good risk management skills (i.e., know what the heck you're doing).
Conclusion
Leveraging Treasuries is yet another tool you can use to beat the public at trading. It pairs particularly well with having broad exposure to equities, as Treasuries tend to shine the most when the economy is weaker than expected, which is exactly when you need liquidity the most from bear markets in stocks. When combined with volatility targeting, risk parity, and factor models for stocks, bonds, commodities, and mortgage-backed securities, you can get about 3x-4x the return of stocks for the same risk.
