Investing Expertise

Can researchers train machine learning models to mimic top mutual fund managers? In his August 2024 paper entitled “Machine Learning from the Best: Predicting the Holdings of Top Mutual Funds”, Jean-Paul van Brakel seeks to anticipate and exploit the stock picking of top-performing U.S. equity mutual fund managers by:

• Using a large set of holdings-based measures and component firm/stock factor exposures and accounting ratios to rank stocks on the likelihood that top managers will pick them. He considers five models: three machine learning models (decision tree, random forest and gradient boosting), a linear (logit regression) model and simple selection based on recent holdings. For machine learning models, he each quarter uses quarterly data from four years ago to one year ago for training and quarterly data from last year for generating holdings likelihoods for next quarter.
• Calculating average absolute Shapley value for each firm/stock characteristic to assess its importance in the decision process.
• Assessing economic value of predictions for each model by each quarter creating six groups of stocks, first sorting stocks into large and small and then sorting each size sort into thirds based on likelihood that top fund managers will pick the stocks. He then each quarter reforms a probable-minus-improbable (PMI) factor portfolio that is long large and small stocks with the highest likelihoods and short those with the lowest.

He applies a 6-factor (market, size, book-to-market, profitability, investment, momentum) model based on daily calculations to compare alphas across funds, features and the PMI factor portfolio. Using quarterly/daily data for U.S. broad equity mutual fund holdings/returns and associated individual firm/stock data starting December 1998 and ending December 2022 and December 2023, respectively, he finds that: Keep Reading

Are large and sophisticated investors (whales) better than small retail investors (minnows) at timing established crypto-asset markets? In their August 2024 paper entitled “Beneath the Crypto Currents: The Hidden Effect of Crypto ‘Whales'”, Alan Chernoff and Julapa Jagtiani compare short-term timing abilities of whales and minnows trading Ethereum (ETH). Specifically, they explore relationships between next-day ETH returns and ETH holdings in e-wallets of four size groups: (1) more than $1 million (whales); (2) $100,000 to $1 million; (3) $10,000 to $100,000; and, (4) less than $10,000 (minnows). They control for supply of ETH in circulation and major crypto-asset market events. Using daily data for ETH from Coin Metrics, including price (midnight to midnight) and holdings/value by e-wallet size group, during January 2018 through December 2023, they find that:

Keep Reading

Do analysts systematically ignore the connection between future firm earnings and current economic conditions? In their July 2024 paper entitled “Predicting Analysts’ S&P 500 Earnings Forecast Errors and Stock Market Returns Using Macroeconomic Data and Nowcasts”, Steven Sharpe and Antonio Gil de Rubio Cruz examine the quality of bottom-up forecasts of near-term S&P 500 earnings aggregated from analyst forecasts across individual firms. Specifically, they:

• Model expected aggregate S&P 500 quarterly earnings growth as a function of GDP growth, output and wage inflation and change in dollar exchange rate. They also consider a simplified model based only on real GDP growth and change in the dollar exchange rate.
• Calculate the gap between modeled S&P 500 earnings growth and analyst-forecasted growth.
• Estimate how well this forecast gap predicts analyst forecast errors.
• Test the extent to which the forecast gap predicts S&P 500 Index total returns.

Using quarterly actual and forecasted S&P 500 earnings, S&P 500 Index total return and values for the specified economic variables during 1993 through 2023, they find that: Keep Reading

Do active investment managers as a group successfully time the stock market? The National Association of Active Investment Managers (NAAIM) is an association of registered investment advisors. “NAAIM member firms who are active money managers are asked each week to provide a number which represents their overall equity exposure at the market close on a specific day of the week (usually Wednesday). Responses can vary widely [200% Leveraged Short; 100% Fully Short; 0% (100% Cash or Hedged to Market Neutral); 100% Fully Invested; 200% Leveraged Long].” The association each week releases (usually on Thursday) the average position of survey respondents as the NAAIM Exposure Index (NEI).” Using historical weekly survey data and Thursday-to-Thursday weekly dividend-adjusted returns for SPDR S&P 500 (SPY) over the period July 2006 through late July 2024, we find that: Keep Reading

Does Michael Farr, CEO and founder of Farr, Miller & Washington, offer good stock picks via his annual CNBC articles identifying the best 10 stocks for the next year? To investigate, we take his picks for 2022, 2023 and 2024, calculate the associated annual (first half only for 2024) total returns for each stock and compute the equal-weighted average return for the 10 stocks for each year. We use SPDR S&P 500 ETF Trust (SPY) as a benchmark for these averages. Using year-end or end of June 2024 dividend-adjusted stock prices for the specified stocks-year, we find that: Keep Reading

Does subjective technical analysis truly add value? In their June 2024 paper entitled “The Power Of Price Action Reading”, Carlo Zarattini and Marios Stamatoudis investigate the value added to simple trading rules by the discretionary judgments of an experienced technician for a sample of stocks with: (1) overnight gaps up over 6%; (2) minimum opening price $2.00; and, (3) minimum pre-market volume at least 200,000 shares. One of the paper’s authors (Marios Stamatoudis) is the expert technician. They assess his abilities both to select the best gaps to trade and to micromanage precise entry points, stop-losses and partial exits at predetermined profit points. To screen out confounding information, they remove dates, ticker symbols, sectors/industries, news, and specific prices/volumes from the his inputs, leaving only an anonymized visual 2-year daily price history for each stock. They present gaps to him randomly, not in chronological order. Using daily pre-gap prices and 1-minute post-gap prices for NYSE and NASDAQ stocks satisfying the above three criteria during January 2016 through December 2023 (9,794 events), they find that:

Keep Reading

Can artificial intelligence (AI), in the form of a deep-learning neural network, outperform human analysts in predicting stock prices at an annual horizon? In the May 2024 revision of his paper entitled “Can AI Replace Stock Analysts? Evidence from Deep Learning Financial Statements”, Nathan Dong tests the ability of a 5-layer neural network to predict stock prices 12 months hence using recent financial statements, interest rates and stock prices. He does not impose any valuation formulas (such as discounted cash flow) or valuation ratios on the AI model. Instead, he lets the neural network, running on inexpensive computer equipment, devise its own approach. He splits his dataset dynamically into an in-sample training subset and an out-of-sample prediction subset. For example, to make a 12-month prediction in the first quarter of 2022, the training subset consists of 5,000 iterations based on the four quarterly financial statements from 2020 and interest rates/stock prices from 3, 6, 9 and 12 months ago. He measures AI model and human analyst accuracy based on mean squared error between target prices and actual future prices and the ability of variation in target prices to explain variation in actual future prices (R-squared). He also explores the factors that drive similarities/differences in target prices between human analysts and the AI model. Using quarterly financial statements, analyst 12-month target prices and actual prices for a broad sample of U.S. stocks as available, and contemporaneous yields for  10-year constant maturity U.S. Treasury notes and Moody’s seasoned Baa corporate bonds during 2010 through 2023, he finds that:

Keep Reading

Do funds based on holdings/trades of members of the U.S. Congress and their families beat the market? To investigate, we look at performances of two recently introduced exchange-traded funds (ETF):

1. Unusual Whales Subversive Democratic Trading ETF (NANC) – invests primarily in stocks held by Democratic members of Congress and/or their families per public disclosure filings.
2. Unusual Whales Subversive Republican Trading ETF (KRUZ) – invests primarily in stocks held by Republican members of Congress and/or their families per public disclosure filings.

We use SPDR S&P 500 ETF Trust (SPY) as a benchmark. Using daily dividend-adjusted prices for NANC, KRUZ and SPY during February 7, 2023 (NANC and KRUZ inception) through June 7, 2024, we find that: Keep Reading

Are large language models such as GPT-4 as effective as professional human analysts in interpreting numerical financial statements? In their May 2024 paper entitled “Financial Statement Analysis with Large Language Models”, Alex Kim, Maximilian Muhn and Valeri Nikolaev investigate whether GPT-4 can analyze standardized, anonymized financial statements to forecast direction and magnitude (large, moderate or small) of changes in future firm earnings and provide the level of confidence in its answer. They withhold management discussions that accompany financial statements, choosing to evaluate the ability of GPT-4 to analyze only numerical data. They anonymize statements by omitting firm names and replacing years with labels (t, t − 1, …) so that GPT-4 cannot use its training data to find actual future earnings. They consider both a simple query and a series of prompts designed to make GPT-4 think like an ideal human analyst by focusing on changes in certain financial statement items, computing financial ratios and generating economic interpretations of these ratios. They compare GPT-4 forecasts to: (1) consensus (median) human earnings forecasts issued during the month after financial statement release; and, (2) forecasts from other benchmarks, including that of a highly focused state-of-the-art artificial neural net (ANN) model. To test economic value of forecasts, they each year on June 30 form portfolios using GPT-4 forecasts based on annual financial statements from the preceding calendar year end, as follows: 

• Sort stocks based on GPT earnings forecasts.
• Select stocks expected to have moderate/large increases or decreases in earnings and separately resort these two groups based on forecast confidence.
• Form an equal-weighted or value-weighted long (short) portfolio of the tenth, or decile, of these stocks with highest confidence in earnings increases (decreases).

Using financial statements for 15,401 firms during  1968 through 2023 (with 2022 and 2023 out-of-sample with respect to the GPT-4 training period), annual returns of associated stocks and consensus human analyst earnings forecasts for 3,152 firms during 1983 through 2021, they find that: Keep Reading

Can large language models like ChatGPT work with numbers to support technical analysis of stock returns rather than just words to support sentiment analysis? In his April 2024 paper entitled “StockGPT: A GenAI Model for Stock Prediction and Trading”, Dat Mai introduces StockGPT, an autoregressive model trained and tested on stock returns rather than firm news. He segments 1926 through 2000 daily U.S. stock returns into intervals (tokens) and then trains StockGPT to identify predictive return patterns. He then tests the model, using daily returns over the past 256 trading days to predict next-day returns during 2001 through 2023. He assesses accuracy of return predictions by:

1. Running cross sectional regressions of actual versus predicted daily returns.
2. Each day at the market close: (1) removing stocks in the lowest tenth of market values; and, (2) reforming an equal-weighted or value-weighted hedge portfolio that is long (short) stocks within the highest (lowest) tenth, or decile, of predicted returns.

He performs sensitivity tests that account for portfolio costs, portfolio construction delay and elimination of low-priced stocks. Using daily returns for all common stocks traded on NYSE, AMEX or NASDAQ during 1926 through 2023, he finds that: Keep Reading