Systematic Fixed Income Investing: How Rules-Based Bond Strategies Work
Systematic fixed income uses written rules to pick duration, credit and bonds. A Treasury trend rule through 2022 and 2023 shows where it helps and where it breaks.
Systematic fixed income investing means picking bonds with written rules instead of a manager's judgment. The rules decide duration, credit exposure and when to step aside. Not a forecast. A process.
Equity investors have used rules-based factors for decades. Bond investors came later. The bond market trades over the counter, data is messy, and many bonds trade rarely. That's changed over the last ten years. The approach now shows up in index funds, model portfolios and institutional mandates.
What is systematic fixed income investing?
It's any bond strategy where a model, not a person, makes the repeatable decisions. A person still writes the rules and checks the data. After that, the rules run.
Three questions drive almost every bond portfolio. How much interest rate risk to hold. How much credit risk to hold. Which specific bonds to own. A systematic process answers each one with a measurable signal.
A discretionary manager answers them too. The difference is consistency. A rule makes the same call on the same data every time. You can test it on history before you trust it.
What factors work in bonds?
Research on corporate bonds points to four. Israel, Palhares and Richardson laid them out in a 2018 paper on common factors in corporate bond returns.
- Carry. Bonds that pay more yield, after adjusting for risk, tend to earn more. It's the bond version of a high dividend.
- Value. A bond whose spread is wide compared with its default risk looks cheap. The model buys cheap spreads.
- Momentum. Bonds, and the issuer's stock, that have done well lately tend to keep doing well for a while.
- Defensive or quality. Lower leverage, steadier profits and shorter maturities. These bonds have often delivered better returns per unit of risk.
Government bonds have their own version. Asness, Moskowitz and Pedersen found value and momentum in bond markets across countries in 2013. Trend signals on Treasury futures are a core part of most managed futures programs.
If you know factor investing in stocks, the logic carries over. The data work doesn't.
How does a systematic bond strategy work, step by step?
- Define the universe. For example, investment-grade U.S. corporates with enough trading volume to buy and sell.
- Score every bond. Each factor gets a score. Spread versus default risk for value. Recent return for momentum. Leverage for quality.
- Combine the scores. Most programs blend factors so one bad year in one signal doesn't sink the portfolio.
- Set limits. Match the benchmark's duration, cap any issuer, cap any sector. The tilt comes from bond choice, not a hidden rate bet.
- Trade on a schedule. Monthly or quarterly. Bond trading costs are high, so turnover rules matter more than in stocks.
The same idea works at a broader level. A rule can set how much duration to hold, or when to move into Treasury bills.
Worked example: a trend rule on Treasuries, 2021 to 2024
Here's a simple rule anyone can check. At each month-end, compare an index tracker for 7-10 year Treasuries with its own 10-month average price. Above the average, hold it next month. Below, hold 1-3 month Treasury bills.
We measured it with adjusted closes on index-tracking funds. These are hypothetical results with no fees or trading costs. They're not the record of any THOR product.
- 2021. The rule held Treasuries only three months. It lost 2.6 percent. Buy and hold lost 3.3 percent.
- 2022. The 10-year yield rose from 1.51 percent to 3.88 percent. The rule held Treasuries for one month and lost 0.7 percent. Buy and hold lost 15.2 percent.
- 2023. Rates swung both ways. The rule got whipsawed, buying late and selling late. It lost 2.9 percent while buy and hold gained 3.6 percent.
- 2024. Another choppy year. The rule lost 1.9 percent. Buy and hold lost 0.6 percent.
Across 2021 to 2023 the rule lost about 6 percent. Buy and hold lost about 15 percent. Almost all of that gap came from one year.
For context, 2022 was the worst year in modern bond history. The U.S. aggregate bond index fell 13.0 percent. Long Treasuries fell 31.2 percent. Investment-grade corporates fell 17.9 percent. Treasury bills gained 1.4 percent.
When does systematic fixed income fail?
This is the part most product pages skip. It's also where the decisions get made.
Choppy rate markets. 2023 and 2024 show it. Trend rules need a sustained move. When yields reverse every few months, the rule pays twice, once getting out and once getting back in.
Trading costs. A corporate bond can cost far more to trade than a large-cap stock. A factor that looks strong on paper can vanish after costs. Many published bond factor results assume prices nobody could trade at.
Bad data. Many bonds don't trade daily. Prices come from dealer quotes or pricing models. A momentum signal built on stale prices measures the pricing service, not the market.
Crowding and liquidity. In March 2020, credit spreads blew out and bond ETFs traded below their net asset value. Every model selling the same bonds at once found few buyers. Rules don't create liquidity.
Short history. Clean bond data covers fewer decades than stock data. Fewer independent cycles means more room for a backtest to fit noise.
Systematic vs active fixed income: what's the difference?
Active bond managers make calls on rates, credit and individual issuers. Some are excellent. Results depend on the people, and the reasoning can be hard to audit.
A systematic manager publishes the logic, or at least can explain it fully. The trade-off is flexibility. A rule can't read a Fed statement or spot a fraud in a footnote. Many firms now blend both, using models to screen and people to veto.
For a broader look at how bond funds hold and price their bonds, see how bond ETFs work. For the cash side of the trade, see capital preservation strategy.
How does THOR use systematic rules?
THOR's strategies run on written rules that move between risk assets and cash-like holdings. The rules seek to reduce drawdowns. They follow the same logic as the Treasury example above, and they carry the same weakness in choppy markets. A rule that steps aside early will sometimes step aside for nothing.
Definitions
- Systematic fixed income: a bond strategy where written, testable rules make the repeatable decisions.
- Duration: a bond's price sensitivity to a 1 percentage point change in interest rates.
- Credit spread: the extra yield a bond pays over a Treasury of similar maturity.
- Carry: the return a bond earns if yields and spreads don't change.
- Bond momentum: the tendency of recent bond winners to keep outperforming for a period.
- Trend rule: a signal that holds an asset above its moving average and steps aside below it.
- Whipsaw: a loss from a signal that flips back and forth in a market without a clear direction.
From January 2021 to December 2023, the 10-month rule held 7-10 year Treasuries in just 9 of 36 months.
Hypothetical results are shown for illustration only. They don't reflect fees, taxes or trading costs. Past performance is not indicative of future results.
More Insights
What Is Risk Parity? How Equal-Risk Portfolios Work, and the Year They Broke
Risk parity builds a portfolio where every asset contributes the same amount of risk. Stocks are volatile, so they get a small weight. Bonds are calm, so they get a large one, and the manager usually borrows to bring the return back up.
Portfolio Risk Management Software: What It Measures and Where It Fails
Risk software measures how much a portfolio could lose. It doesn't act. A 60/40 stress test through 2008, 2020 and 2022 shows where the models break.
Factor Investing: A Systematic Approach, From Fama-French to Machine Learning
Factor investing is a systematic approach to picking stocks by a trait you can measure. How the rule runs, what Andrew Ang and machine learning add, and the dated stretches where factors failed.
Get The Signal Every Morning
Brad Roth's daily market brief — systematic signals, ETF positioning, and what the data is actually showing. Free to subscribe.
Subscribe to The Signal