Investing Expertise

Can large language models (LLMs) create financial research? In their December 2024 paper entitled “AI-Powered (Finance) Scholarship”, Robert Novy-Marx and Mihail Velikov describe a process for automatically generating academic finance papers using LLMs and demonstrates its efficacy by producing hundreds of complete papers on stock return predictability. Specifically, they:

1. Identify 31,460 potential stock return predictors from accounting variables and their differences.
2. Screen these potential signals for redundancy, data robustness and stock selection breadth to identify 17,074 candidates for validation.
3. Validate these signals via decile and quintile portfolio sorts and controls for multiple known stock return factors to select 183 promising (alpha) signals.
4. Apply a series of anomaly evaluation tools to isolate 96 economically meaningful and statistically reliable signals and generate a standardized report for each that details signal performance, including a comparison to over 200 other known anomalies.
5. Use state-of-the-art LLMs to generate three conventional academic papers for each effective signal (288 reports in total), including:
   • Creative signal names.
   • Abstract.
   • Introductions with motivation, hypothesis development, results summary and contribution.
   • Data and conclusion.
   • Citations to other relevant papers.

Based on results from this process, they conclude that: Keep Reading

How can investors best apply the available array of Large Language Models (LLM) in quantitative strategy development? In his December 2024 paper entitled “The LLM Quant Revolution: From ChatGPT to Wall Street”, William Mann summarizes the use of LLMs in quantitative finance, focusing on: the current state of LLM technology in financial applications; comparison of leading models; implementation frameworks for production systems and risk management; and, quality control considerations. He reports results of a comprehensive review of the best LLM to use for each of seven phases of investment strategy development. Based on the body of research on use of LLMs in finance and this comprehensive review, he concludes that: Keep Reading

Are the human choices in studies that apply machine learning models to forecast stock returns critical to findings? In other words, is there a confounding machine learning design choices zoo? In their November 2024 paper entitled “Design Choices, Machine Learning, and the Cross-section of Stock Returns”, Minghui Chen, Matthias Hanauer and Tobias Kalsbach analyze effects of varying seven key machine learning design choices: (1) machine learning model used, (2) target variable/evaluation metric, (3) target variable transformation (continuous or discrete dummy), (4) whether to use anomaly inputs from pre-publication subperiods or not, (5) whether to compress correlated features, (6) whether to sue a rolling or expanding training window and (7) whether to include micro stocks in the training sample. They examine all possible combinations of these choices, resulting in 1,056 machine learning models. For each machine learning model each month, they:

1. Rank stocks on each of 207 potential return predictors and map rankings into [-1, 1] intervals. In case of missing inputs, they set the ranking value to 0.
2. Apply rankings to predict a next-month target variable (return in excess of the risk-free rate, market-adjusted return or 1-factor model risk-adjusted return) for each stock with market capitalization above a 20% NYSE threshold during January 1987 through December 2021.
3. Reform a hedge portfolio that is long (short) the value-weighted tenth, or decile, of stocks with the highest (lowest) predicted target variable and compute next-month portfolio return.

Using monthly data as available for all listed U.S. common stocks during January 1957 through December 2021, they find that: Keep Reading

How should investors think about potential shocks  to firm valuations and financial markets research from generative artificial intelligence (GenAI)? In their October 2024 paper entitled “AI and Finance”, Andrea Eisfeldt and Gregor Schubert review the literature on the effects of GenAI on (1) firm valuations and (2) financial research methods. They also offer an introduction to available GenAI research tools and advice on using these tools. Based on the body of research, they conclude that:

Keep Reading

Despite access to elaborate trading platforms and real-time data, the large majority of speculative traders incur substantial losses (see, for example the chart below). In his August 2024 paper entitled “The Myth of Profitable Day Trading: What Separates the Winners from the Losers?”, Franklin Gallegos-Erazo identifies factors that distinguish the few successful traders from the many who fail, including risk management, emotional control and strategies employed. Based on results of past studies, he concludes that: Keep Reading

Is detailed decision-making prowess a better metric than past performance for comparing portfolio managers? In their October 2024 paper entitled “Actions Speak Louder Than (Past) Performance: The Relationship Between Professional Investors’ Decision-Making Skill and Portfolio Returns”, Isaac Kelleher-Unger, Clare Levy and Chris Woodcock examine the link between professional investor decision-making and overall performance for long-only stock portfolios involving at least 80 decisions per year. Specifically, they analyze daily positions for each stock to quantify seven decision outcomes: stock-picking, entry timing, scaling in, size adjusting, weighting, scaling out and exit timing. They then aggregate effects of all decisions at the portfolio level relative to prospectus benchmarks or, where none is stated, to a relevant index. They measure added values of decision types as follows (see the figure below):

1. Stock picking – positive or negative overall return to the position while owned.
2. Entry timing – proximity of initial entry price to its low from 21 trading days before through 21 trading days after purchase.
3. Scaling in – comparison of return to a buy-and-hold strategy at average price of the stock from initial entry to first sell trade.
4. Adding/trimming/no-trade – comparison of return to buy-and-hold at the median quantity from first sell trade to the first sell trade after the last add trade.
5. Scaling out – comparison of return to a buy-and-hold strategy at average price of the stock from the first sell trade after the last add trade to the total exit.
6. Position weighting – comparison of return to that for a hypothetical equal-weighted portfolio.
7. Exit timing – proximity of final exit price to its high from 21 trading days before through 21 trading days after purchase.

They then combine hit rate (fraction of decisions with positive value-add) and payoff ratio (ratio of value-add to value-loss across all decisions) for each investor to compute a Behavioral Alpha (BA) Score, and relate BA Score to current and future portfolio performance. Using proprietary daily holdings of 123 long-only stock portfolios managed by professional investors during 2013 through 2023, they find that:

Keep Reading

How do exchange-traded-funds (ETF) that employ artificial intelligence (AI) to pick assets perform? To investigate, we consider ten such ETFs, eight of which are currently available:

• • Amplify AI Powered Equity ETF (AIEQ) – picks U.S. stocks based on a quantitative model that runs on the IBM Watson platform.
  • QRAFT AI-Enhanced U.S. Large Cap ETF (QRFT) – picks large capitalization U.S. stocks via an investment strategy enhanced by the use of AI.
  • QRAFT AI-Enhanced U.S. Large Cap Momentum ETF (AMOM) – picks momentum stocks via an investment strategy enhanced by the use of AI.
  • VanEck Social Sentiment ETF (BUZZ) – picks U.S. stocks with the most positive sentiment as determined from online sources via a rules-based quantitative methodology.
  • WisdomTree International Al Enhanced Value Fund (AIVI) – picks developed market value stocks outside the U.S. and Canada via a quantitative AI model. [We use data only since retention of an AI-oriented advisor at the beginning of 2022.]
  • WisdomTree U.S. Al Enhanced Value Fund (AIVL) – picks U.S. value stocks via a quantitative AI model. [We use data only since retention of an AI-oriented advisor at the beginning of 2022.]
  • Teucrium Aila Long-Short Agriculture Strategy ETF (OAIA) – takes long and short positions in agricultural securities via a strategy employing machine learning. [Dead as of August 2024.]
  • Teucrium Aila Long-Short Base Metals Strategy ETF (OAIB) – takes long and short positions in base metals securities via a strategy employing machine learning. [Dead as of May 2024.]
  • LG QRAFT AI-Powered U.S. Large Cap Core ETF (LQAI) – – picks large capitalization U.S. stocks via an investment strategy enhanced by the use of AI.
  • Stocksnips AI-Powered Sentiment US ALL Cap ETF (NEWZ) – picks U.S. stocks by leveraging AI to generate News Media Sentiment Signals.

We use SPDR S&P 500 ETF Trust (SPY) for comparison, though it is not conceptually matched to some of the ETFs. We focus on monthly return statistics, along with compound annual growth rates (CAGR) and maximum drawdowns (MaxDD). Using monthly total returns for the ten AI-powered ETFs and SPY as available through October 2024, we find that: Keep Reading

Since 1990, the Federal Reserve Bank of Philadelphia has conducted a quarterly Survey of Professional Forecasters. The American Statistical Association and the National Bureau of Economic Research conducted the survey from 1968-1989. Among other things, the survey solicits from experts probabilities of U.S. economic recession (negative GDP growth) during each of the next four quarters. The survey report release schedule is mid-quarter. For example, the release date of the third quarter 2024 report is August 9, 2024, with forecasts through the third quarter of 2025. The “Anxious Index” is the probability of recession during the next quarter. Are these forecasts meaningful for future U.S. stock market returns? Rather than relate the probability of recession to stock market returns, we instead relate one minus the probability of recession (the probability of good times). If forecasts are accurate, a relatively high (low) forecasted probability of good times should indicate a relatively strong (weak) stock market. Using survey results and quarterly S&P 500 Index levels (on survey release dates as available, and mid-quarter before availability of release dates) from the fourth quarter of 1968 through the third quarter of 2024 (224 surveys), we find that:

Keep Reading

Do some mutual fund managers game Morningstar ratings/benchmarks by shifting the styles of their funds? In their September 2024 paper entitled “Box Jumping: Portfolio Recompositions to Achieve Higher Morningstar Ratings”, Lauren Cohen, David Kim and Eric So investigate how mutual fund managers exploit investor reliance on Morningstar ratings by adjusting holdings to jump their funds into size/value styles with low benchmarks. They focus on active U.S. and global equity mutual funds during the period from five years before to five years after June 2002, when Morningstar began rating funds by style. They include dead funds to avoid survivorship bias. Using Morningstar style assignments, Morningstar ratings and performance data for active equity mutual funds during 1997 through 2007 (with some data through 2022), they find that:

Keep Reading

Can Google’s large language model (LLM), Gemini, beat simple benchmarks by picking a small portfolio of stocks? In their September 2024 paper entitled “Can AI Beat a Naive Portfolio? An Experiment with Anonymized Data”, Marcelo Perlin, Cristian Foguesatto, Fernanda Müller and Marcelo Righi test the ability of Gemini 1.5 Flash to weight a portfolio of five U.S. stocks for different investment horizons (1, 6, 12 or 36 months) using either financial data, price data or a combination of both. The tests employ real but anonymized financial data, ensuring that confounding training data about specific firms does not infect stock weighting decisions. Testing involves 18,000 iterations of five steps:

• Select a random date during 2004 through 2023 with equal daily probability.
• Based on this date, randomly select five U.S firms/stocks that have financial data, prices over $1.00 and average daily volumes over $250,000 over the prior five years.
• Create anonymized versions of the financial and price data.
• Ask Gemini to assign allocations from a fixed total amount to the five stocks, with the possibility of also allocating to the contemporaneous 5-year U.S. Treasury note yield.
• Compare performance of the Gemini-weighted portfolio during the given investment horizon to that of the S&P 500 Index and an equal-weighted portfolio of the same stocks/Treasury note yield.

Using inputs as specified for 1,522 distinct firms during 2004 through 2023, they find that: Keep Reading