Can Machine Learning Predict Factor Returns?
Can machine learning techniques improve the prediction of cross-sectional factor returns in equity markets?
Can machine learning techniques improve the prediction of cross-sectional factor returns in equity markets?
As portfolios incorporate more sustainability data—from climate impact assessments to labor practices and board diversity metrics—a critical question emerges: Does this wealth of ESG information actually enhance portfolio performance, or is it merely additional data without tangible investment value?
In the ongoing debates about the virtues of active versus passive investment strategies, a fundamental problem undermines meaningful discussion: there is no universally accepted definition of what constitutes a "passive" strategy.
Letdin, Seagraves, and Sirmans advanced our understanding of REIT asset pricing by developing and rigorously testing six REIT-specific return factors—size, value, momentum, earnings quality, low volatility, and short-term reversal—using decades of data.
Handy and Meksi provided a clear warning: relying on past performance to select CTAs is a strategy fraught with risk, largely due to behavioral biases that distort our perception of skill and persistence. For investors, the lesson is to stay humble, diversify, and focus on robust processes rather than chasing yesterday’s winners.
Fueled by the persistent failure of active management, passive investing now commands the majority of assets under management.
Is volatility (the standard deviation of returns) a good measure of the risk that investors actually care about?
The empirical research (for example, here, here, here and here) on insider trading demonstrates that insider transactions have significant predictive power for future stock returns [...]
Momentum investing remains a viable strategy. However, the way you construct and manage your momentum portfolio matters greatly.
Christian Goulding and Campbell Harvey, authors of the study "Investment Base Pairs," proposed a groundbreaking framework for portfolio construction that challenges traditional approaches in modern finance. Their research focused on leveraging cross-asset information to optimize investment strategies and improve returns across diverse asset classes. Here's an overview of their investigation, key findings, and takeaways for investors and advisors.
Profitability subsumes all of the quality factor, explaining both the performance of the strategies the investment industry market and the factors that academics employ—none of the quality factors generated significant positive alpha relative to profitability, the other Fama and French factors, and momentum.
The financial research literature has found that the performance of assets (and factors) can vary substantially across regimes - factor premiums can be regime dependent. Unfortunately, the real-time identification of the current economic regime is one of the biggest challenges in finance.
The main benefit of constructing industry momentum portfolios based on standard ICS is that it is straightforward and reproducible. However, that benefit may come at the cost of accuracy and oversimplification of complex industry relationships between companies.
For equity investors there have been two major narratives over the last 17 calendar year period 2008-2024. The first is that US stocks have far outperformed international stocks. The other narrative has been the outperformance of growth stocks relative to value stocks.
The empirical research we have reviewed shows that the (hidden) costs of index construction and rebalancing policies to investors are about 10 times the expense ratios.
This paper provides an introductory overview of infrastructure investing, exploring its characteristics, benefits, challenges, and potential role in a diversified portfolio.
US exceptionalism provided the same explanation for the outperformance of US stocks in the 1990s. However, that regime changed. From 2000-2007, while the S&P 500 Index returned just 1.9% per annum (underperforming riskless one-month Treasury bills by 1.3% per annum), the MSCI EAFE Index returned 5.6% per annum, and the MSCI Emerging Markets Index returned 15.3% per annum.
The listing domicile explained about 50% of the valuation gap. In other words, US-listed stocks are substantially more expensive than internationally listed stocks for no reason other than the place of listing.
NAV timing investors could potentially create trading strategies which would systematically transfer wealth from buy-and-hold investors to themselves.
While the media headlines are preaching doom, the fundamentals are telling a very different story—credit spreads have widened, and EBITDA multiples are the lowest they have been in a decade. The bottom line is that for investors able to accept its limited liquidity, private, senior, secured and sponsored by private equity direct lending continues to be a compelling component of a diversified portfolio deliver what has always attracted investors: high current income, resilience through market cycles, and a disciplined approach to risk management. We are far from a bubble.
© Copyright 2025 alpha architect | All Rights Reserved | Home | Terms of Use | Privacy Policy | Disclosures | Subscribe | Contact Us
