Recent Academic Research
New alpha in crypto options, REITs, and directional trading. Also, LLM time series forecasting and aging affects on investment ability.
This week, I’m sharing five recent academic papers on financial markets that are too good to miss. Just a heads-up: this post is a bit strategy-heavy, but I promise it’s worth it.
Oh, and, as always, everything here is completely free. If you know someone who’d love some fresh, unbiased, data-driven market insights, feel free to share this with them. Let’s get into it.
Crypto Options
Exploiting Inflows Into Ethereum with Options
Paper Title: Return-forecasting and Volatility-forecasting Power of On-chain Activities in the Cryptocurrency Market
Authors & Date: Yeguang Chi, Qionghua (Ruihua) Chu, & Wenyan Hao - 11/10/2024
Summary:
This paper investigates the connection between on-chain activity, particularly ETH net inflows into cryptocurrency exchanges, and the returns and volatility of ETH. The authors find a statistically significant negative relationship between ETH net inflows and future ETH returns. This suggests that when a large volume of ETH flows into exchanges, it signals a movement to sell, creating a bearish market sentiment that drives prices down.
Additionally, high ETH net inflows are associated with lower ETH volatilities. A key takeaway for traders is that this bearish sentiment can be exploited by selling 0DTE ETH call options when net inflows are high, which proves to be a profitable strategy with a win rate of at least 81.82% in various scenarios.
Thoughts:
First crypto paper on the newsletter. I’m pretty neutral on crypto myself, but it’s wild how divisive this asset class is. Pro-crypto, anti-crypto... the debates are endless. Maybe it’s the new-kid-on-the-block factor?
Anyway, this paper caught my eye because it outlines a successful crypto options trading strategy. And if you know anything about options, you’ll know selling a naked call is no joke—it’s basically daring the market to ruin your day (unlimited losses if the crypto price skyrockets).
With Bitcoin sitting at over $90,000 right now, betting against the market for the long haul sounds like a terrible idea. Thankfully, this strategy uses super short intraday time frames—from 1 to 6 hours.
Interestingly, the magic doesn’t seem to work for BTC, which suggests this alpha might be fading fast. If you trade ETH options, though, it could be worth exploring before the edge disappears.
Commercial Mortgage-Backed Securities and REITs
Relative Value Trading - REITs and CMBX Index
Paper Title: 15 Seconds to Alpha: Higher Frequency Risk Pricing for Commercial Real Estate Securities
Authors & Date: Andreas Christopoulos & Joshua Barratt - 11/13/2024
Summary:
This paper introduces a model for estimating CMBX risk partitions in 15-second intervals and uses it to develop profitable trading strategies. These strategies exploit the relationship between these risk signals and REIT pricing to generate buy and sell signals for REITs.
By fusing the CMBX risk signals with REIT pricing in 24 distinct trading strategies, the researchers achieved statistically significant alphas in 88% (21/24) of the strategies. Furthermore, 90% (19/21) of the significant alpha strategies generated positive cumulative returns ranging from 9.09% to 41.37% during the first year of the COVID-19 pandemic. The Sharpe ratios for these successful strategies ranged from approximately 2 to 5.
The study also found that the liquidity and excess liquidity risk partitions were the most valuable and reliable indicators for successful trading strategies. The success of these strategies was consistent across the entire sample period, encompassing various credit rating classes.
Thoughts:
CMBX and REITs might be niche, but this paper proves that overlooked markets can offer huge opportunities. The use of 15-second interval signals is a testament to how technology is reshaping trading, especially in extracting alpha from micro-inefficiencies. Liquidity signals were the clear winners here, reinforcing how liquidity—or the lack of it—drives returns during market stress.
And those Sharpe ratios? Anything above 2 is exceptional, but 5 is almost unheard of, signaling not just high profitability but also low portfolio standard deviation. For anyone hunting untapped edges, this research is a masterclass in blending data granularity with robust strategy design.
Large Language Models
LLMs Are Better at Forecasting
Paper Title: Quantifying Qualitative Insights: Leveraging LLMs to Market Predict
Authors & Date: Hoyoung Lee, Youngsoo Choi, & Yuhee Kwon - 11/13/2024
Summary:
This paper investigates the use of LLMs to forecast stock market movements based on qualitative insights from securities firm reports. The authors develop a method for quantifying qualitative factors and integrating them with numerical data. To evaluate the effectiveness of this approach, the researchers compared the performance of LLMs, specifically LLaMA2 and GPT-4-Turbo, to traditional time-series models, ARIMA and LSTM.
The results showed that the LLMs consistently outperformed the baseline models in predicting market movements, achieving higher accuracy and Matthews correlation coefficient (MCC) scores, particularly when considering a one-day look-back period. This improved performance suggests that LLMs can better incorporate qualitative information and capture nuances in market trends.
Thoughts:
AI and LLMs research papers are everywhere these days, but this paper is a standout. The fact that an LLM outperformed established time series models in forecasting stock prices highlights how rapidly these models are evolving. From my anecdotal experience, on the buy-side, adoption varies—some firms are diving in, while others remain cautious.
As LLMs grow more accurate, we could see a significant shift away from traditional statistical methods for market analysis. The real question is how quickly the edge from these models gets arbitraged away once they go mainstream. For now, they offer an interesting glimpse into the future of predictive analytics in finance.
Trading Directional Changes
Eight Directional Trading Strategies
Paper Title: Optimization of Multi-Threshold Trading Strategies in the Directional Changes Paradigm
Authors & Date: Ozgur Salman, Themistoklis Melissourgos, & Michael Kampouridis - 11/4/2024
Summary:
The sources describe a trading model, MSTGAM, that leverages the Directional Changes (DC) paradigm for stock market prediction. The DC paradigm analyzes price movements based on significant changes, defined as movements exceeding a predefined threshold (θ). These significant changes mark DC events, while price fluctuations within a trend that surpass the threshold but move against the trend are considered Overshoot (OS) events.
The MSTGAM incorporates eight trading strategies: two based on scaling laws observed in the DC paradigm and six that utilize indicators derived from DC and OS events. Notably, the model employs multiple thresholds and a genetic algorithm to optimize the weights assigned to different strategies and thresholds, ultimately aiming for profit maximization.
The MSTGAM demonstrated significantly better performance compared to traditional technical analysis strategies and market indices with a high Sharpe Ratio of 5.59 and a 22% rate of return. These results underscore the potential of the DC paradigm and the effectiveness of the MSTGAM in generating profits while mitigating risk in stock market trading.
Thoughts:
Now, this is another optimization paper, so you need to take the strategy success with a grain of salt. However, in my time in finance, I had yet to hear about the Directional Changes paradigm.
The approach could be particularly intriguing for those relying on historical price patterns to design strategies, as it offers a fresh perspective on market behavior. There’s a lot to unpack here, and if you’re into innovative frameworks for trading, this paper is definitely worth a deep dive.
Behavioral Economics
Old Age and Worse Performance
Paper Title: Older Investors at a Loss: Cognitive Aging and Funds Returns
Authors & Date: Zhongtai Li, Jia Liu, & Yanran Wu - 11/12/2024
Summary:
The authors of the paper investigated the effects of cognitive aging on the returns of individual investors in mutual funds using data on Chinese investors from 2006 to 2011. They found that cognitive aging negatively affects investor returns, supporting their hypothesis that cognitive aging is negatively correlated with returns. This negative relationship between age and returns followed an inverted U-shape, with the right tail flattening, meaning that cognitive aging does have a negative impact, but the impact lessens as investors reach retirement age.
The authors found that retirement can mitigate the negative effects of cognitive aging, possibly because retirees have more time to manage their investments and process information efficiently. The paper concluded that cognitive aging is a factor that should be considered when making investment decisions.
Thoughts:
You might say “But Alpha in Academia, of course one should consider the age and mental capability of a fund manager before investing with them!” And while I completely agree with that statement, the mental decline of an individual may not be all that visible, which is why it is important to understand the implications from this paper.
The authors highlight how retirement helps mitigate the effects of cognitive aging, possibly due to having more time to analyze investments and strategies. But let’s not overlook other factors: reduced stress, a shift in priorities, or even the removal of cognitive biases that come with managing others’ money. It’s a fascinating reminder that performance isn’t just about strategy—it’s also about the mind behind it.
Feedback
Thank you for reading this week’s edition of Recent Academic Research. Remember to fill out the poll to let me know which paper was your favorite. Feel free to follow up with any questions, comments, or ideas for the future!
Also, make sure to follow me on X / Twitter!
Disclaimer
The content provided in this newsletter, "Alpha in Academia," is for informational and educational purposes only. It should not be construed as financial advice, investment recommendations, or an offer or solicitation to buy or sell any securities or financial instruments. Past performance is not indicative of future results. The financial markets involve risks, and readers should conduct their own research and consult with qualified financial advisors before making any investment decisions.
The interpretations, opinions, and analyses presented herein are those of the author and do not necessarily reflect the views of the original researchers, their institutions, or the full implications of the cited academic papers. While every effort is made to accurately represent the research discussed, readers should be aware that the summaries and interpretations may not capture the full scope or nuances of the original studies. The information contained in this newsletter is believed to be accurate and reliable at the time of publication, but accuracy and completeness cannot be guaranteed. The author and publisher accept no liability for any loss or damage resulting from reliance on the information provided.
This newsletter may contain links to external websites or resources. The author is not responsible for the content, accuracy, or reliability of these external sources.
By subscribing to and reading this newsletter, you acknowledge that you have read and understood this disclaimer and agree to hold the author and publisher harmless from any liability that may arise from your use of the information contained herein.
Among the various papers you referred to, the one on the forecasting capacity of LLMs caught my attention. Not only for the research itself; but also because it is potentially linked to various papers that concern how AI positively (or negatively) influences decision making in risk, investment or "high-stakes" scenario situations. A paper that I found very useful and interesting in this sense is: https://journals.sagepub.com/doi/abs/10.1177/17456916231181102