Today i want to build some factor model which is not commonly used but has significant…
- User
Today i want to build some factor model which is not commonly used but has significant outperformance. Can you recommend me some factor model that is not on usual fama, capm +give their original papers and how they define it
Assistant
Of course. This is an excellent goal. Moving beyond the canonical Fama-French and CAPM models into more recent or specialized factor models is where much of the interesting work in quantitative finance is happening.
Here are three academically rigorous and powerful factor models that are less "mainstream" than the standard ones, along with their original papers and construction details. These have all demonstrated significant historical outperformance and are built on strong economic or behavioral intuitions.
1. The q-Factor Model (Investment and Profitability)
This is arguably the most significant challenger to the Fama-French 5-Factor model. It is grounded in economic theory (specifically, investment theory) rather than being purely empirically motivated.
- Core Idea: The model argues that a company's investment decisions and its profitability are the primary drivers of expected returns, alongside market risk and size.
- Original Paper: Hou, K., Xue, C., and Zhang, L. (2015). Digesting Anomalies: An Investment Approach. The Review of Financial Studies, 28(3), 650-705.
Factor Definitions:
The q-factor model explains stock returns using four factors:
- MKT (Market Factor): Same as in all other models. The excess return of the value-weighted market portfolio over the risk-free rate.
- ME (Size Factor): Same as Fama-French's SMB. The return of small-cap stocks minus the return of large-cap stocks.
- I/A (Investment Factor): This is the key innovation. It captures the "investment anomaly" where companies that invest more conservatively outperform companies that invest aggressively.
- How it's defined: Stocks are sorted based on their Investment-to-Assets (I/A) ratio. I/A is the annual change in total assets divided by the lagged total assets.
- Construction: The portfolio is constructed by going long stocks with low I/A (conservative investment) and short stocks with high I/A (aggressive investment). The intuition from investment theory is that managers will only invest in new projects if the expected return meets a certain hurdle rate (the cost of capital). High investment implies a low cost of capital (and thus low expected returns), while low investment implies a high cost of capital (and high expected returns).
- ROE (Profitability Factor): This is the second key innovation, capturing profitability.
- How it's defined: Stocks are sorted based on their Return on Equity (ROE). ROE is calculated as income before extraordinary items divided by lagged book equity.
- Construction: The portfolio is constructed by going long stocks with high ROE (high profitability) and short stocks with low ROE (low profitability). The intuition is straightforward: more profitable firms are better investments and should command higher expected returns.
Why it's compelling: The authors of the q-factor model show that it can explain a vast array of market anomalies that the Fama-French 3-Factor and Carhart 4-Factor models struggle with. Its factors are also grounded in valuation theory, making it intellectually appealing.
2. Quality Minus Junk (QMJ)
While "Quality" is a term often used loosely, AQR Capital Management's principals (Asness, Frazzini, and Pedersen) created a specific, rigorous, and academically-tested definition for it.
- Core Idea: High-quality stocks—those that are safe, profitable, growing, and well-managed—consistently deliver high risk-adjusted returns. Conversely, low-quality ("junk") stocks deliver poor risk-adjusted returns.
- Original Paper: Asness, C. S., Frazzini, A., and Pedersen, L. H. (2019). Quality Minus Junk. The Review of Financial Studies, 32(1), 34-72.
Factor Definition:
The QMJ factor is not based on a single metric but is a composite of several "quality" dimensions.
-
The "Quality Score": A stock's quality is determined by its aggregate score across several categories. For each metric within a category, stocks are ranked and assigned a z-score (how many standard deviations they are from the mean). These z-scores are then averaged to create the final quality score.
-
The Components of Quality:
- Profitability: The most important dimension. Measured by metrics like Gross Profit / Assets, Return on Equity (ROE), Return on Assets (ROA), and Cash Flow / Assets. The goal is to identify firms that generate high and stable profits.
- Growth: Measures the growth of profitability over the past few years. The idea is to reward consistent growth, not just a one-time profit spike.
- Safety / Stability: This is about risk. Measured by metrics like low beta, low idiosyncratic volatility, low earnings volatility, and low financial leverage. The idea is that safer, more stable companies are higher quality.
-
Construction of the QMJ Factor:
- At each rebalancing date (e.g., monthly), all stocks are ranked by their composite quality score.
- The factor portfolio is constructed by going long the top 30% of stocks with the highest quality scores and short the bottom 30% of stocks with the lowest quality scores ("junk").
- The portfolio is typically value-weighted and rebalanced periodically.
Why it's compelling: The QMJ factor has shown strong, consistent performance globally and across different asset classes. The intuition is that investors may systematically underprice quality (perhaps because it's "boring") and overpay for speculative, lottery-like junk stocks, creating a persistent premium for holding quality.
3. Betting Against Beta (BAB)
This factor model frontally attacks one of the core assumptions of the CAPM—that higher beta (systematic risk) should be rewarded with higher returns. The evidence shows this relationship is flat, or even negative, on a risk-adjusted basis.
- Core Idea: A portfolio that is long low-beta stocks and short high-beta stocks, constructed to be market-neutral, generates significant positive returns. This is because leverage-constrained investors bid up the prices of high-beta assets, depressing their future returns.
- Original Paper: Frazzini, A., and Pedersen, L. H. (2014). Betting Against Beta. Journal of Financial Economics, 111(1), 1-25.
Factor Definition:
The construction of the BAB factor is more complex than a simple long/short portfolio because it involves targeting a specific portfolio beta.
-
Rank by Beta: All stocks in the universe are ranked based on their historical beta, estimated over a period like the past year of daily returns.
-
Form Two Portfolios:
- Low-Beta Portfolio: A portfolio of all stocks with below-median betas.
- High-Beta Portfolio: A portfolio of all stocks with above-median betas.
- Within each portfolio, stocks are typically weighted by the inverse of their beta rank (so the lowest-beta stocks get the most weight in the low-beta portfolio, and the highest-beta stocks get the most weight in the high-beta portfolio).
-
Leverage and De-leverage to Beta 1: This is the critical step.
- The Low-Beta Portfolio is leveraged (by borrowing at the risk-free rate) so that its ex-ante beta is exactly 1.0. For example, if the portfolio's beta is 0.6, you would hold 1.67 of the portfolio for every 0.67.
- The High-Beta Portfolio is de-leveraged (by shorting the stocks and investing a portion of the proceeds in the risk-free asset) so that its ex-ante beta is also exactly 1.0. For example, if the portfolio's beta is 1.5, you would hold 0.67 in the short portfolio and keep 1 of capital.
-
Construct the BAB Factor:
- The BAB factor return is the return of the long, beta-adjusted low-beta portfolio minus the return of the short, beta-adjusted high-beta portfolio.
Why it's compelling: The BAB factor exploits a powerful behavioral and structural anomaly. Many investors (like traditional mutual funds) are prohibited from using leverage. To chase higher returns, they are forced to buy high-beta stocks, pushing their prices too high and lowering their risk-adjusted returns. Sophisticated investors can exploit this by longing leveraged low-beta assets and shorting these overpriced high-beta assets.
Practical Considerations for Building These Models:
- Data: You will need access to high-quality historical data, including market data (prices, volumes) from a source like CRSP and fundamental company data (assets, equity, income) from a source like Compustat.
- Implementation Details: The "devil is in the details." Pay close attention to rebalancing frequency, how you handle delistings, value-weighting vs. equal-weighting, and the specific definitions of metrics (e.g., how to calculate beta or assets).
- Transaction Costs: The published papers report gross returns. In reality, transaction costs can significantly eat into the profits of any strategy, especially those with higher turnover.
- Factor Decay: Be aware that as factors become more well-known and traded, their premia can shrink over time (a phenomenon known as "factor crowding"). However, the three models above are based on robust economic and behavioral rationales, which suggests they may be more persistent than purely data-mined anomalies.
Good luck with your model building! These three provide excellent, well-documented starting points for exploring the frontiers of factor investing.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.