Weekly Research Recap
Latest research on investing and trading
This week’s recap highlights the most useful investing insights from academia, industry reports, blogs, and social media, with every source linked.
Crypto
The MicroStrategy Trap: Why Your “Bitcoin Proxy” is Actually a Leveraged Bet (Krause)
MicroStrategy operates less like a Bitcoin proxy and more like a leveraged corporate vehicle. Its structure amplifies even modest crypto moves: A recent 17% Bitcoin decline triggered a 48% drop in MSTR, and stress tests signal similar downside convexity. With a 0.62 correlation to Bitcoin and a 2.4 beta to equities, the exposure is far from pure. Key takeaway: Investors seeking Bitcoin returns should use spot ETFs while MSTR adds uncompensated leverage and corporate fragility.
Quantitative Evaluation of Volatility-Adaptive Trend-Following Models in Cryptocurrency Markets (Karassavidis, Kateris, and Ioannidis)
A trend-following model for BTC and ETH delivers strong and stable performance. After optimizing ATR-scaled position sizing, RSI-based momentum confirmation, and SMA trend rules across 5 million parameter sets, the final configuration produces Sharpe ratios of 1.54 (BTC) and 1.80 (ETH), with CAGRs of 20% and 24% and maximum drawdowns near 11%. Key takeaway: Well-designed trend-following strategies remain a source of risk-adjusted returns in crypto.
Timing Usage of Technical Analysis in the Cryptocurrency Market (Zatwarnicki and Zatwarnicki)
The paper introduces the Rolling Strategy–Hold Ratio (RSHR), a rolling-window method that evaluates any strategy from every possible start date. Across roughly 50,000 simulations, simple TA rules underperform in traditional markets but add value in alternative ones, while sentiment models, especially Fear & Greed, show the most stable performance. Key takeaway: Strategy results are highly time-window dependent, making rolling evaluation essential to avoid false confidence.
Equities
How Investors Pick Stocks: Global Evidence from 1,540 AI-Driven Field Interviews (Hwang, Noh, and Shin)
How do investors pick stocks? Interviews suggest they buy stocks based on strong financials, clear product traction, brand familiarity, expert cues, volume-confirmed momentum, dividend needs, and simple survival checks like “Will this firm still exist in five years?” Most combine 2–3 such rules, varying across countries. Key takeaway: Stock picking is governed by practical, heterogeneous heuristics, many of which are not captured by standard asset-pricing models.
Theme Washing in the ETF Space (He, Wang, and Ward)
Roughly a quarter of thematic ETF–month observations “theme wash,” overstating their true thematic exposure. These funds earn an alpha shortfall of −2.5% per year, yet still attract flows because investors respond to persuasive thematic language rather than actual holdings. Matching the marketing language of genuine thematic ETFs also allows managers to charge higher fees and softens outflows after poor performance. Key takeaway: In thematic ETFs, narrative often substitutes for substance; pay attention to portfolio holdings, not the pitch.
Stock Valuation and Investor Expectations (Danielson)
The paper reviews key stock-valuation methods, constant-growth, multi-stage, and expectations-based models, and shows how each helps investors interpret the growth implied by current prices. It demonstrates that high valuation ratios often require 10–30 years of sustained competitive advantage or 20–30%+ returns on new projects, assumptions few firms can deliver. Key takeaway: High valuations imply demanding growth bets; investors must judge whether those expectations are realistic.
Hedge Funds
On Evaluating the Style-Selection Skill of Hedge Funds (Ye, Li, and Tee)
A subset of hedge-fund managers consistently tilt toward styles that generate higher risk-adjusted alpha in the following months, a strategic skill distinct from stock picking or market timing. These high-skill funds outperform low-skill peers by 0.94% per month with correspondingly higher alpha, show forecasting power for up to 12 months, and face significantly lower liquidation risk. Key takeaway: Style selection is a persistent and economically meaningful source of hedge-fund outperformance.
Investing
A Refinement to the Treynor Ratio (Brzeszczynski, Gajdka, Pietraszewski, and Schabek)
The authors show that ranking funds by the Treynor Ratio, excess return per unit of market risk (beta), breaks down whenever excess returns or betas are negative, producing systematic misrankings. In a 982-fund sample, roughly 5% exhibit negative betas, enough to generate large distortions. Their Modified Treynor Ratio eliminates these anomalies and restores economically consistent rankings. Key takeaway: Use the refined metric to avoid misinterpreting fund performance.
Machine Learning and Large Language Models
What Drives the Performance of Machine Learning Factor Strategies? (Esakia and Goltz)
Machine-learning factor models deliver strong gains in stylized tests, but most vanish under realistic conditions. In ideal settings, broader information adds about 1.2%/month and nonlinearities about 1.0%/month. Once microcaps, hindsight, and trading costs are removed, nonlinearity adds little, and short exposure becomes essential. Key takeaway: Real-world ML alpha comes from information breadth rather than complexity, and mostly requires a long–short implementation.
Generative AI for Analysts (Xue, Zhang, and Zhu)
AI tools push analysts to produce broader, denser research, drawing on many more inputs, covering wider topics, and using a larger analytical toolkit, while enabling faster report turnaround. Yet forecast precision deteriorates sharply as errors rise and analysts struggle to compress rich information into a single earnings number. Key takeaway: AI boosts the scale and speed of research, but at the expense of forecast accuracy.
Volatility
Combining realized volatility estimators based on economic performance (Skintzi and Fameliti)
The paper shows that volatility forecasts become far more useful when combined using economic payoffs rather than statistical errors. Weighting models by their past portfolio utility, straddle-trading profits, or VaR accuracy consistently beats individual RV measures and standard combinations. These improvements remain robust across crises, volatility regimes, horizons, and forecast models. Key takeaway: Volatility forecasts are most valuable when combined using economic objectives, not statistical fit.
Blogs
Patience Pays: Why Quality Shares Outperform in the Long Run (CFA Institute)
Alternative Market Signals: Investing with the Box Manufacturing Index (Quantpedia)
Murphy’s Law (Quantitativo)
Podcasts
Rob Hanna - Trading the VIX in a Diversified Portfolio (The Algorithmic Advantage)
This Trader Turned $3K Into $1,000,000.. Here’s How (Words of Rizdom)
The $20 Trillion Absurdity | Jim Grant on the Risk No One Sees (Excess Returns)
Social Media & Industry Research
Gold and Bitcoin (Campbell Harvey)
Opportunity-Set Bias in Mean-Reversion Trading Systems (Carlo Zarattini)
The Shape of Fear: Managing Risk Through Options Markets (Man Group)
Goldman’s $2 Billion ETF Deal Exposes Shareholder Blind Spots (Aaron Brown)
Last Week’s Most Popular Link
The Best Defensive Strategies: Two Centuries of Evidence (Baltussen, Martens, and van der Linden)
Forecast-Agnostic Portfolios (Guo and Wachter)
Agentic Artificial Intelligence in Finance: A Comprehensive Survey (Aldridge et al.)
Disclaimer: This newsletter is for informational and educational purposes only and should not be construed as investment advice. The author does not endorse or recommend any specific securities or investments. While information is gathered from sources believed to be reliable, there is no guarantee of its accuracy, completeness, or correctness.
This content does not constitute personalized financial, legal, or investment advice and may not be suitable for your individual circumstances. Investing carries risks, and past performance does not guarantee future results. The author and affiliates may hold positions in securities discussed, and these holdings may change at any time without prior notification.
The author is not affiliated with, sponsored by, or endorsed by any of the companies, organizations, or entities mentioned in this newsletter. Any references to specific companies or entities are for informational purposes only.
The brief summaries and descriptions of research papers and articles provided in this newsletter should not be considered definitive or comprehensive representations of the original works. Readers are encouraged to refer to the original sources for complete and authoritative information.
This newsletter may contain links to external websites and resources. The inclusion of these links does not imply endorsement of the content, products, services, or views expressed on these third-party sites. The author is not responsible for the accuracy, legality, or content of external sites or for that of any subsequent links. Users access these links at their own risk.
The author assumes no liability for losses or damages arising from the use of this content. By accessing, reading, or using this newsletter, you acknowledge and agree to the terms outlined in this disclaimer.