Weekly Research Recap
Latest research on investing and trading
Here’s your latest weekly briefing featuring the most actionable investing insights from the past seven days, spanning academic research, industry reports, blogs, and social media, complete with direct links throughout.
Asset Allocation
Using monthly data from 1979 to 2025, simple equity trend filters (10-month MA or 12–1 momentum) preserve equity-like returns (about 10% CAGR) while slashing maximum drawdowns from −54% to about −20% and lifting Sharpe from 0.72 to 0.93. Key takeaway: Trend following should be viewed more as dynamic risk control rather than alpha, materially reducing crash risk while preserving equity-like long-run returns.
Crypto
Systematic Trend-Following with Adaptive Portfolio Construction: Enhancing Risk-Adjusted Alpha in Cryptocurrency Markets (Bui and Nguyen)
AdaptiveTrend combines 6-hour trend signals with volatility-scaled trailing stops, monthly Sharpe-filtered asset selection, and a 70/30 long–short tilt. Across 150+ crypto pairs, it delivers 40.5% annualized return, Sharpe 2.41, and a −12.7% max drawdown, significantly beating TSMOM and buy-and-hold. Dynamic trailing stops contribute roughly +0.7 Sharpe. Key takeaway: In crypto, intermediate-frequency trends plus adaptive risk and portfolio construction can turn raw momentum into robust performance.
Equities
Risk Appetite and (Mis)Pricing (Guo, Li, Li, and Li)
Using a risk appetite index, the paper shows that the CAPM works only when aggregate risk aversion is high. In those periods, the security market line (SML) is strongly upward sloping (226 bps per unit of beta per month) with an insignificant intercept. When risk aversion is low, the SML flips negative (slope of −77 bps) and high-minus-low beta alphas reach −153 bps/month. Key takeaway: The beta anomaly is state-dependent, and risk appetite governs whether beta earns a premium or reflects mispricing.
Expected Investment Growth and REIT Returns (Liang, Eshraghi, and Wang)
Expected future investment growth predicts returns across U.S. REITs. A long–short portfolio (long high-growth, short low-growth) earns about 0.51% per month (6% annually) and remains significant after standard factor controls. High-growth REITs systematically increase leverage and their cash-flow risk, explaining the risk premium. Key takeaway: Forward-looking growth expectations are a priced signal in REITs that investors can potentially harness.
Multiples for Valuation: Go High, Go Low, Ignore the Middle (Estrada)
Multiples predict 10-year returns mainly when valuations are extreme. Using U.S. data from 1871 to 2025, dividend, earnings, and CAPE yields show much higher in-sample correlations and out-of-sample forecast accuracy in the top/bottom quartiles (CAPE yield correlation up to 0.70), while predictive power largely fades in the middle range.
Key takeaway: Long-horizon valuation signals matter most at extremes.
Value vs. Growth: What Drives the Value Premium? (Chen, Huang, and Jiang)
About half of the value and growth stocks are newly classified each year. The New Value–New Growth spread earns 0.33% per month versus 0.15% for incumbents, with a higher alpha, Sharpe, and a much fatter right tail, driven mainly by the persistent underperformance of new growth stocks. Key takeaway: A disproportionate share of the value premium comes from recent style migrants, especially newly promoted growth stocks.
Factor Models
Covariance Implied Risk Factors (Kaebi)
Standard PCA distorts latent factors when assets exhibit heterogeneous idiosyncratic variances, overweighting high-variance assets. Heteroskedastic PCA fixes this by iteratively correcting the diagonal of the covariance matrix, lifting out-of-sample Sharpe ratios by roughly 50 to 150%. Key takeaway: Improve factor estimation by correcting heteroskedasticity before expanding the factor set.
Machine Learning and Large Language Models
High-Frequency Trading in the Chinese Stock Market: Predictability, Profitability, and Regulation (Wang, Wang, and Zhou)
Ultra-high-frequency returns in China are strongly predictable: Machine-learning models achieve 15.5% out-of-sample R² at 5-second horizons and about 60% directional accuracy, stronger than comparable U.S. results. Predictability spikes near daily price limits and is driven mainly by order-flow imbalance and price-limit proximity. Key takeaway: Millisecond alpha exists, but it’s fragile and weakens once arbitrage constraints ease.
This is a survey paper discussing how electricity price forecasting is shifting toward deep learning with probabilistic outputs. Day-ahead models increasingly use Transformers and GNNs, while intraday research is moving toward orderbook-based signals. Key takeaway: Forecasting performance in electricity markets depends as much on market-specific model design as on the choice of architecture.
Options
The Impact of Early Option Exercise on Ex-Dividend Stock Returns (Sperling and Schlie)
Early exercise of in-the-money calls before ex-dividend dates triggers mechanical selling on the ex-day. Across 270k U.S. dividend events, a 1-σ rise in exercise activity lowers ex-day returns by 6.4 bps, while top-decile events underperform by 12 to 16 bps. The effect strengthens with high dividend yields and weakens when options are liquid. Key takeaway: Early option exercise amplifies ex-dividend day selling pressure.
Trading
Improving Performance with Fast Alphas; A Tactical Overlay for Intraday Trend Trading (Zarattini and Pagani)
Fast intraday signals can look exceptional in frictionless backtests (a 5-minute reversal delivers 32% gross CAGR with Sharpe >2) but turn unprofitable after transaction costs. Used instead as an execution overlay on a slower intraday trend strategy, the same signal adds 200 bps per year and lifts Sharpe from 0.87 to 0.99 by improving entry and exit timing. Key takeaway: Fast signals that fail standalone after costs can still add meaningful value through smarter execution.
Blogs
Big Picture Asset Pricing (John H. Cochrane)
Seeing Through the Shutdown’s Missing Inflation Data (New York Fed)
Volatility Clustering Across Asset Classes: GARCH and EGARCH Analysis with Python (2015–2026) (Jonathan Kinlay)
1951 (John H. Cochrane)
Podcasts
Quality of Earnings and Dilution Risk with Ryan Telford (Planet MicroCap)
You Can’t Eat Risk-Adjusted Returns | AQR’s Pete Hecht on Portable Alpha’s Capital Efficient Edge (Excess Returns)
“I think of everything as a bet” - Ex-SIG Quant Trader Andrew Courtney (Odds on Open)
The Cost-Benefit of Being Trendy ft. Andrew Beer & Tom Wrobel (Top Traders Unplugged)
Social Media & Industry Research
A Trend Following Deep Dive: AlphaTrend and Agentic Research Workflows (Man Group)
Why Bonds Still Belong: Rethinking Fixed Income in Modern Portfolios (Return Stacked, Corey Hoffstein)
A Visual Appreciation For Black-Scholes Delta (Kris Abdelmessih)
Last Week’s Most Popular Links
The Unpriced Risk in Momentum Strategies (Gao and Yuan)
QuantaAlpha: An Evolutionary Framework for LLM-Driven Alpha Mining (Han et al.)
Pairs Trading with Time-Series Deep Learning Models (Yilmaz and Sefer)
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.
Paid subscriptions may not be available in all jurisdictions and may change without notice.