Inspirations

Today, another interesting strategy from the Monthly Mover category, this time on US bonds. We explore the observation that this instrument owes practically all its returns to the last part of the month.

From the perspective of an international investor, the American debt market seems quite complicated, mainly due to the specific terminology of bonds.

The entire American debt market is referred to as "Bonds." However, it's important to know about its internal division into three categories:

• Short-term, called Treasury Bills (T-Bills), with a maturity of less than a year. A popular ETF is TBIL (3M)

• Medium-term, called Treasury Notes (T-Notes), with a maturity of 2 to 10 years. A popular ETF is IEF (7-10Y)

• Long-term, called Treasury Bonds (T-Bonds), with a maturity of 20 to 30 years. Example ETFs are TLT (20Y)
  

Today, we will focus on a strategy exploring a pattern in long-term bonds, specifically on TLT.

Why bother with bonds at all? For decades, they have been cited as an example of a stabilizing instrument for a stock portfolio, i.e., an instrument that is lowly correlated or even inversely correlated with stock prices. Just read about the 60/40 portfolio.

We won't delve into the details of treating bonds as an investment here. Statistically, it is an investment with a much lower return rate than stocks, and additionally, as shown in the years 2020-2023, this stable haven turned out to be a nightmare for many financial institutions (a price drop of over 48%), contributing to the banking crisis in the USA at the beginning of 2023.

However, since we are looking for instruments and strategies that are lowly or inversely correlated with stocks, today we are tackling TLT. Let's get to work!

Is it true that there is a negative correlation between TLT and S&P500 (SPY)?

The chart below shows that in the long term, YES.

Throughout most of the period from 2003 to 2023, bonds were inversely correlated with stocks (an increase in one was accompanied by a decrease in the other). The year 2022 was exceptional because stocks and long-term bonds fell simultaneously, which was painfully felt by investors using the 60/40 model. The average correlation (the purple line on the bottom chart) places us in the middle today, indicating a lack of correlation, which is precisely what we aim for.

Bond Monthly Mover withstood the challenges of this instrument from 2020 to 2024 and closed this very negative period for bonds with profits.

Key Components of the strategy

• Exploration of the monthly pattern on TLT.

• The strategy includes two simple filters to improve its effectiveness.

  
  

Backtest 1, Fixed $ Money Management

In this variant, we consistently invest the same amount of $100k. We test the period of the last 22 years for the years 07.2002 - 02.2025.

• Invested capital: $100,000

• Test period in years: 22

• Tested years: 07.2002 - 02.2025

• Tested Instrument: TLT

Equity Chart for this test:

Basic statistics and results month by month:

Examples of transactions on the chart:

Click the button to see the latest backtest:

Backtest 2, % Money Management

In this backtest, we are investing in a strategy that constantly uses 100% of the current capital (starting with $100'000 capital). This means that as the capital grows or decreases, the value of the position changes proportionally. The rest of the parameters remain unchanged.

The equity chart for this test looks as follows (yellow line is a benchmark).

Basic statistics resulting from the test:

Trading Strategy Analysis

Net Profit and CAGR

A net profit above $148,030 in the studied strategy is significantly lower than the Benchmark (Buy & Hold SPY marked in yellow on the chart), which amounts to $856,827, translating to a CAGR of 4.22% vs 10.81%. This indicates that the studied strategy achieves significantly lower net profit and a lower average annual return rate, suggesting its lesser effectiveness in generating profits over the long term.

Drawdown and Return/Drawdown Ratio

The maximum open drawdown in the analyzed strategy was -8.36%, compared to -55.19% in the benchmark. This results in a much better return/open drawdown ratio of 8.13 versus 4.75. This indicates that the analyzed strategy is less risky and more stable.

Exposure

The exposure in the analyzed strategy was only 13.9% vs 100% in the benchmark. This means that the analyzed strategy was "in the market" only 14% of the time. This results in less market risk exposure, and the Risk Adjusted Return was approximately 20%. You can read more about exposure here.

Winning Percent

The winning percent in the analyzed strategy was 60%. This means that this percentage of transactions was profitable, which is a fairly comfortable result. Additionally, the average profit was 1.3 times greater than the average loss (Win/Loss ratio).

SL & TP

The strategy does not use a typical stop loss and take profit, although there are no obstacles to introducing them. According to our tests, for most strategies on ETFs/stocks, these settings worsen results (see why). The protection against a very strong impact of a potential price change of a single stock on the entire portfolio is the diversification within the portfolio of various strategies. The strategy uses a Time Exit. Visit the stop loss order page.

Market Regime

The strategy has been tested in all basic market regimes and includes filters implemented based on this. You can find more on this topic here.

Trading Costs

Trading costs and slippage were taken into account in the backtests. You can check our last research about trading costs using Alpaca Broker here. With a diversified portfolio of stocks and strategies, transaction costs can determine your profit or loss, so take the time to thoroughly test and choose a broker.

Robustness

The number of historical transactions 170, is significantly lower than Stockpicker-type strategies, which can provide even over 100,000 transactions in robustness tests. We adhere to the principle that the fewer parameters, the greater the strategy's robustness. Therefore, we strive to ensure our strategies have as few parameters as possible and to select only those parameters that significantly impact the strategy's effectiveness while aligning with its nature.

Recommended Instruments

The recommended primary instrument for this strategy on Algocloud is TLT, which has shown the best historical results.

• Primary Instrument: TLT
  

Pattern Day Trader

The strategy statistically did not close any trades on the same day, so it does not meet the criteria of a Pattern Day Trader (PDT). This means it can also be used on smaller accounts.

Correlation

The strategy is inversely correlated with most other strategies, which helps balance a stock strategy portfolio. This is a significant advantage. To check the strategy's correlation with others, visit the correlations page.

Summary & Strengths and Weaknesses of the strategy

Strengths of the strategy:

• Pattern stability. Over 22 years, the studied strategy had only 5 losing years, at the very beginning of the TLT instrument's listing, which is a very good result for this type of strategy despite the very turbulent recent years for TLT.

• Inverse correlation to other strategies. The correlation with most other strategies is negative, meaning it should balance well with a stock strategy portfolio.

• Low drawdown.  The Max Open Drawdown in the analyzed strategy was only 8.36% compared to 55.19% in the Benchmark.

• Positive winning percent and Win/Loss Ratio indicators. 60% of transactions ended with a profit, and the average profit was 1.3 times greater than the average loss, highlighting the potential convenience of using the strategy.

• No same-day transactions. A plus of the strategy is its applicability to smaller accounts.

• Low capital commitment. The strategy offers relatively low capital commitment with an exposure of 13.9% vs. Benchmark's 100%, meaning it can be successfully used in a portfolio alongside other strategies.
  

Weaknesses of the strategy:

• Profitability. In the analyzed strategy, the Net Profit amounted to $148,030, while the Benchmark achieved $856,827. The CAGR was 4.22%, which is lower than the 10.81% for the Benchmark. It should be noted that due to low exposure, the Risk Adjusted Return was approximately 20%.

• Robustness. The strategy involves fewer transactions compared to other Stockpicker-type strategies, which can be seen as a disadvantage in terms of robustness. As a result, it needs closer monitoring and should be integrated into a portfolio with more robust, diverse strategies.
  

Summary

Bonds are not and never will be growth leaders. Compared to other stock strategies, the annual return rate of this strategy is not overwhelming. We have a return of just under 4% annually, but with capital used for only about 14% of the time, which results in an 20% annualized return.

The main advantage of this strategy, however, is the inverse correlation with stocks that bonds statistically provide, as seen in the test results. Although the decline in bond and stock prices in 2022 was exceptionally parallel, as longer history shows, this is not the rule but rather an exception. Unlike bonds alone, the strategy emerged defensively from this period, showing all turbulent years in the positive, and thanks to the inverse correlation with stocks, it can provide stability to a portfolio.

If you want to include bonds in your portfolio, this strategy is definitely worth considering as an alternative to simply buying and holding this instrument.

What you get in the package for this strategy:

• .SQX file ready to use on the Algocloud and StrategyQuant platforms.

• Pseudocode that describes all the rules in an easy-to-understand way.