026 - Gary Antonacci II - New Research & Updated Models

Absolute Trend, Rotation Between Models & Assets, & 100 Year Back-Tests

A new research paper titled "A Century of Profitable Industry Trends" by Zarattini (Concretum Research) and Antonacci explores the profitability of a long-only trend-following strategy applied to industry portfolios over nearly 100 years, from 1926 to 2024.

The strategy utilizes industry-based trend-following mechanisms, particularly focusing on identifying trends in 48 different industries. The paper distinguishes between two key forms of momentum—time-series momentum and cross-sectional momentum: Time-Series Momentum focuses on the continuation of an asset’s trend, i.e., whether a price continues in the same direction. Cross-Sectional Momentum compares relative performance across assets, betting on the strongest performers and avoiding or shorting weak performers.

The strategy has an impressive long-term track record, averaging an annual return of 18.2% with 12.6% volatility and a Sharpe Ratio of 1.39. In contrast, the U.S. equity market returned 9.7% annually with a Sharpe Ratio of 0.63. The trend-following strategy not only outperforms the market but does so with less risk (reduced volatility and drawdowns).

Risk Management

A critical feature of the strategy is its ability to participate fully during market upswings while minimizing exposure during downturns. The maximum drawdown (the largest peak-to-trough loss) was 33% for the strategy, much lower than the 84% drawdown for a passive market portfolio.

Volatility management is achieved using position-sizing techniques that allocate based on each industry’s recent volatility. The paper suggests limiting leverage to 200% to maximize returns while controlling risks.

Sector ETFs

In the last 20 years, the strategy was applied to 31 sector ETFs, showing that this timing approach could be replicated using readily available financial instruments like ETFs. The authors conducted back-tests to confirm that the same trend-following rules worked effectively with ETFs, even after considering transaction costs and slippage.

Trader Application

For traders, this study highlights a robust and adaptable strategy, providing several key insights:

1. Long-Only Trend Following: The strategy’s success lies in its ability to adapt to market trends by entering long positions when momentum is strong and exiting to avoid downturns. Traders can apply this approach to industries, sectors, or even individual assets showing strong upward trends.

2. Focus on Timing: Entry and exit rules are clearly defined using Keltner Channels and Donchian Channels (a combined approach – whichever is hit first). These technical tools help capture breakout movements while filtering out noise to avoid frequent whipsaws.

3. Risk-Adjusted Profitability: Traders can learn to manage exposure through volatility-targeted position sizing, ensuring that all positions contribute equally to portfolio risk. This method can improve portfolio performance while minimizing overexposure to any single sector or asset.

4. Sector Rotation with ETFs: The back-testing of sector ETFs shows that a trader today can replicate this strategy using ETFs. For example, during periods of economic uncertainty, a trader might rotate into sectors with strong momentum like utilities or technology while avoiding sectors like energy or materials that may be underperforming.

5. Hedging: The strategy emphasizes minimizing risk during downturns by moving into cash or bonds. Traders could augment this with options or futures hedging strategies during periods of market instability to protect gains. Note that stops are implied by exiting when the market hits the lower channel. This stop loss is never adjusted downward, only upward, ensuring the strategy doesn’t lose gains from a winning position.

6. Indicators Used for Entry & Exit: The strategy relies on two technical indicators to capture trends:

- Keltner Channels: These measure price volatility and consist of a moving average line (center) and two outer bands based on the asset's Average True Range (ATR). The upper and lower bands represent breakout points where prices show strength or weakness.

- Donchian Channels: Developed by Richard Donchian, these channels plot the highest high and lowest low over a specific period. When prices cross above the upper band, it signals a breakout to the upside, while crossing below the lower band signals weakness.

Both indicators include a "noise region," allowing prices to oscillate without triggering frequent signals. In this paper the period for the upper band is 20 days, and to force slower exits, the bottom channel is defined by a 40-day period.

7. Universe of Indexes and Instruments

- The model is applied to 48 U.S. industry portfolios from the Kenneth French Data Library, spanning July 1926 to March 2024. These industries are constructed using the 4-digit SIC codes for stocks listed on the NYSE, AMEX, and NASDAQ.

- In the paper's final section, the model is also applied to 31 sector ETFs from State Street Global Advisors over the past 20 years (2005-2024). These ETFs represent a broader universe of sectors, making the strategy applicable to more liquid, real-time instruments.

8. Position Sizing

The strategy employs a volatility-scaling position sizing approach. This means that positions are sized based on the recent volatility of each industry, ensuring that all assets contribute equally to overall portfolio volatility.

- Volatility Target: The target volatility per industry is set at 1.5% divided by the number of assets (N) in the portfolio. If N = 10, then each industry’s target volatility is 0.15%.

- The weight of each asset is inversely proportional to its volatility. Read the paper for the detail!

- Leverage Limits: The portfolio can use leverage up to 200% of its equity. If the total exposure exceeds this limit, all positions are rescaled to meet the leverage constraint.

- Rebalancing: Positions are rebalanced daily to adjust the weights of each asset based on their volatility and market trends.

Conclusions

The timing-based approach outlined in this paper is highly applicable for traders seeking to develop profitable, trend-following strategies in both industry sectors and broader markets. With clearly defined rules and risk management, this approach can deliver consistent outperformance over time.

Naturally you’ll need to tweak this kind of research to make it your own. Use it to conduct research, and from there build a superior model. This is where art meets science in strategy development!

But the Paper Isn’t All We Discussed with Gary

We talked about the updates to his models and his new focus on channel break outs. However, the dual momentum component is still there however I see it as ‘meta-strategy’ logic which rotates capital from one model to another based on relative strength, not just from one asset to another.

We also touched on his mean-reversion trading, which he does with a suite of hand-chosen ETFs, and a bunch more, so tune in to hear the detail.

May the force be with you.

Simon & Rich

Links to the paper & contact details over on the website:

https://thealgorithmicadvantage.com/podcast/026-gary-antonacci/