Factor Analysis: The OGs’ Comeback to Digital Assets

If you hear “factor analysis” your mind might be brought back to business school or CFA reads about the OGs of finance: Fama-French, Carhart, Barra and their corresponding models. While these concepts have for long been well established in the traditional portfolio management and analysis space, they are relatively new in digital assets. In the crypto space, we are seeing portfolio managers and research analysts pay an increasingly high level of attention to factor analysis when evaluating their allocation decisions and when showcasing their portfolio performance to current and prospective Limited Partners (LPs).

Why is this trend taking shape, particularly among research-driven managers adopting fundamental and quantamental investment approaches?

How should factor analysis be used in the digital assets context to produce useful and actionable information for institutional investors?

In this article we delve into this topic, with an additional note on how the advancements in Machine Learning (ML) and Artificial Intelligence (AI) can become particularly useful to make sense of the complexities of digital assets investments.

What is Factor Analysis?

First, a quick recap on what factor analysis actually refers to. Factor analysis is a statistical method used to identify underlying relationships between variables and to explain how different factors contribute to the performance of an investment. In essence, it helps in breaking down the returns of a portfolio into various factors, which can be attributed to specific risks or exposures. This technique helps investors understand the impact of different variables on their portfolio and isolates factors that drive performance.

Why is Factor Analysis Gaining Traction Among Research-Driven Funds?

For research-driven funds that focus on token fundamentals and blockchain ecosystems, factor analysis offers significant advantages. These funds often rely on in-depth research to evaluate the intrinsic value of digital assets based on their technology, use cases, and network effects.

Factor analysis is particularly valuable for these funds in the following ways:

- Dynamic Beta Evaluation: It's crucial to evaluate how betas with respect to certain factors or assets evolve over time. For instance, if a fund aims to maintain a beta of 1 ± 0.1 with respect to Ethereum for one if its portfolios, factor analysis can help identify any drifts from this targeted exposure. By continuously monitoring and analyzing these drifts, portfolio managers can make informed adjustments to their positions, ensuring alignment with their strategic goals.

- Performance Attribution: By breaking down returns into various factors, portfolio managers can attribute performance to specific aspects of their research. This improves transparency and insight into performance for PMs and LPs alike, making it easier to assess which fundamental factors are driving performance.

- Risk Management: Understanding the impact of different fundamental factors on performance helps manage risk more effectively. For example, running retrospective analyses to identify periods where the beta with respect to Ethereum drifted considerably can provide insights into why these deviations occurred, enabling better management and adjustment strategies in the future.

The benefits of robust factor analysis become apparent to portfolio managers both during their investment committee decisions as well as during their LP meetings, where allocators are often interested in gaining more insight into the fund’s research process.

How Can Traditional Factor Models Be Adapted to Digital Assets?

For portfolio managers who prefer using traditional factor models, such as the Fama-French three-factor model, the Carhart four-factor model, or the Barra risk model, adapting these models to the digital assets' world is essential. These traditional models, originally designed for equities, need modifications to account for the unique characteristics of digital assets.

- Fama-French Model: Originally focused on equity market factors like size, value, and market risk, this model can be adapted to digital assets by incorporating factors specific to the cryptocurrency market, such as network activity, technological advancements, or market sentiment indicators. For example, incorporating factors related to blockchain development activity or on-chain metrics could provide a more accurate representation of digital asset performance.

- Carhart Model: The Carhart four-factor model, which includes momentum as an additional factor, can be adapted to digital assets by analyzing momentum within cryptocurrency markets. This might involve examining trends in trading volumes, price momentum, and investor sentiment to capture the behavior of digital assets more effectively.

- Barra Risk Model: The Barra model’s focus on capturing systematic risk factors can be extended to digital assets by including factors related to market liquidity, regulatory news, and technological changes. This helps in understanding the unique risks associated with digital asset investments and allows for more precise risk management.

Incorporating these adaptations is essential for portfolio managers to leverage traditional factor models to better analyze and manage digital asset portfolios.

Can AI/ML Improve Factor Analysis?

Artificial Intelligence (AI) and Machine Learning (ML) can significantly enhance factor analysis for digital assets by providing advanced analytical capabilities and uncovering patterns that traditional methods might miss. AI and ML algorithms can process vast amounts of data, including on-chain metrics, market sentiment, and transaction volumes, to identify emerging factors and refine existing models.

In particular, AI/ML can help with:

‍- Predictive Analytics: AI can improve predictive accuracy by analyzing complex relationships between factors and predicting how changes in one factor might impact returns. Machine learning models can be trained on historical data to forecast future trends and anomalies, offering more dynamic and real-time insights.

- Pattern Recognition: ML algorithms excel at recognizing patterns and correlations in large datasets. This capability can help identify subtle signals in digital asset markets, such as shifts in investor behavior or changes in network activity, that traditional factor models might overlook.‍

- Adaptive Models: AI can enable more adaptive and responsive factor models by continuously learning from new data and adjusting predictions accordingly. This ensures that models remain relevant and accurate as market conditions evolve.

Integrating AI and ML into factor analysis can help portfolio managers gain a deeper and more nuanced understanding of digital asset markets, leading to more informed investment decisions and enhanced risk management.

The Future of Factor Analysis in Digital Asset Investing

The growth in the digital asset industry – from trading volumes to on-chain activity to institutional adoption – will call for increasingly refined and specialized analytical tools. We believe that factor analysis will play a more and more relevant role for investment managers looking to enhance their portfolio and risk management toolkit and navigate the complexities of digital assets.

Reach out to our team at Genie to learn more about applying purpose-built factor analysis frameworks to digital assets and scaling your organization’s investment research capabilities.

‍

About GenieAI:

GenieAI is the premier AI-powered Portfolio and Risk Management System for digital assets, leading the industry towards smarter, data-driven decisions and automated workflows. The technology enables investors to anticipate market trends and navigate risks in ways that were once unimaginable—all while managing portfolios with unprecedented precision.

GenieAI helps leading fund managers to:

1. Enhance portfolio metrics

2. Predict market trends

3. Improve data-driven decisions

4. Automate analysis and reporting

5. Boost model performance

GenieAI's expert team of industry leaders brings unmatched AI and ML experience to investment management. GenieAI’s clients include some of the world’s most successful and well-respected digital asset funds.

GenieAI is backed by leading investors, including Coinbase Ventures, Bain Capital Ventures, Goodwater Capital, Sierra Ventures, Fasanara Digital, Arca, and other prominent firms.