Recent Academic Research
International soccer wins and market performance, information advantage before FOMC meetings, using commodity futures to forecast inflation, and forecasting future corporate bond and S&P 500 returns
Welcome back to another issue of Recent Academic Research! Today, I have three papers on using forecasting models in the financial markets.
Let’s get into it.
Soccer Wins and Market Returns
Paper Title: Sing When You Are Winning: Football Rivalry and Emotional Stock Market Dynamics
Authors & Date: A AlSabah & Ge Gao | 01/31/2025
Summary:
This paper finds that international football (soccer) matches, including World Cup, European Championship, and Copa América games, impact stock markets in 15 countries. When a national team wins, markets see positive returns, while losses lead to declines and increased volatility.
The effect is strongest in rivalry matches, where a win boosts markets more, and a loss causes sharper drops. Playoff matches also lead to lower trading volume and higher post-game volatility, reflecting investor uncertainty. The study highlights how national team performance in major tournaments influences investor sentiment and stock returns.
Thoughts:
As an American (and a huge American football fan), I prefer to call this sport “soccer,” even though the rest of the world may disagree. I find this paper fascinating, as it combines two things that I love (sports and the financial markets).
This is a classic example of how markets are affected by investor emotions that are completely disconnected from the markets. Emotion-based trading in the markets still exists today.
Information Advantage of Prop Firms
Paper Title: The Fed and the Wall Street Put
Authors & Date: Jan Harren, Mete Kilic, & Zhao Zhang | 01/29/2025
Summary:
This paper examines how proprietary trading firms (PTFs) trade S&P 500 options ahead of FOMC announcements, often taking large short positions before the Fed releases policy decisions. Using CFTC Large Trader Reporting (LTR) data and CBOE options trading records, the study finds that PTFs systematically profit from post-FOMC market reactions, especially when the Fed’s decision surprises expectations.
Their trading is most aggressive before meetings with major policy shifts, suggesting either an informational advantage or superior forecasting ability. The findings raise questions about market efficiency and whether some traders have better insight into Fed policy than others.
Thoughts:
We all know that quant firms and proprietary trading firms are full of smart individuals with lots of resources and data. Clearly, they are able to use their information to better forecast future Fed decisions, and consistently make profits off of it.
This is a clear example of alpha being exploited in institutional finance. I don’t think the data sources from this paper are updated frequently enough to try to replicate the trades in real time, but it would be interesting to explore the relationships. I wonder how quickly this edge will be reduced, especially after the publication of this paper.
Forecasting Inflation with Commodities
Paper Title: On the predictive power of food commodity futures prices in forecasting inflation
Authors & Date: Ankush Agarwal, Christian Oliver Ewald, Shuya Zhang, & Yihan Zou | 01/29/2025
Summary:
This study investigates whether food commodity futures prices can improve U.S. inflation forecasts. Using monthly futures data from 1996 to 2023, the authors develop models based on twelve different food commodities, as well as aggregated models using a simple component approach and Principal Component Analysis (PCA).
Their findings indicate that futures-based models, particularly the PCA-based approach, outperform traditional macroeconomic models in forecasting both food inflation and overall CPI inflation across short- and long-term horizons (3 to 12 months). Spot prices show similar predictive power, reinforcing the robustness of these models.
Thoughts:
Many different economists, governmental agencies, and investment funds attempt to forecast and predict future CPI. Now, it is not surprising that food prices (especially prices in the financial markets) are able to have predictive power, given that food prices make up about 14% of CPI.
This goes to show how efficient the market is at determining an accurate and fair price for futures of agricultural commodities. Now, I don’t know if this is enough information alone to construct a great CPI / inflation forecasting model, but it may be a useful part or addition.
Predicting Corporate Bond Returns
Paper Title: Forecasting Corporate Bond Index Returns
Authors & Date: Jie Cao, Linjia Song. Ruijing Yang, & Xintong Zhan | 01/29/2025
Summary:
This paper explores whether corporate bond index returns can be predicted using company fundamentals and macroeconomic trends. The authors analyze 180 firm characteristics (like profitability and leverage) and 65 economic indicators (like inflation and employment) to find useful predictors.
They test different forecasting models and find that Partial Least Squares (PLS) performs best, significantly improving return predictions over traditional methods. Investors using this approach achieve Sharpe ratios of 0.85 (investment-grade bonds) and 0.92 (high-yield bonds), compared to 0.44 for a simple buy-and-hold strategy.
Thoughts:
Another paper on forecasting! Now, this paper uses a lot of different predictor variables, and that makes me concerned about the potential of the model to overfit to the past data. However, the authors due use a train-test split approach to verify the model’s accuracy, and the out-of-sample performance is still quite good.
The trading strategy’s Sharpe ratio is good, but it is by no means amazing. An interesting paper nonetheless.
Forecasting the S&P 500 with Neural Networks
Paper Title: Forecasting S&P 500 Using LSTM Models
Authors & Date: Prashant Pilla & Raji Mekonen | 01/29/2025
Summary:
This paper compares ARIMA and LSTM models for forecasting the S&P 500 Index (SPX). While ARIMA performs well for short-term trends (89.8% accuracy), it struggles with long-term dependencies due to its linear assumptions. LSTM, a deep learning model designed for sequential data, significantly outperforms ARIMA, achieving 96.41% accuracy with much lower prediction errors (MAE: 175.9 vs. 462.1).
Surprisingly, the LSTM model without additional features (like technical indicators and macroeconomic data) performed better than the version with features, suggesting that deep learning can extract relevant patterns from price data alone. The findings reinforce LSTM’s potential for financial market forecasting, with future improvements focusing on hybrid models and advanced deep learning techniques.
Thoughts:
The third forecasting paper today! I promise I did not do this on purpose (these were just the best papers that I found from the past week).
Not a whole lot to say here, as I have featured prior papers on using LSTM to forecast equity returns. This paper agrees with prior papers that LSTM models are more accurate than traditional ARIMA models. This makes me wish my undergraduate class on time series forecasting included LSTM models in addition to ARIMA and GARCH models!
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 and like the post if you enjoyed it.
Feel free to follow up with any questions, comments, or ideas for the future!
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 or 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.