Investing Expertise

Given that finance is ultimately tied to human emotions, does the body of research support belief that abilities of machine learning to handle large amounts of data, non-linearities and variable interactions will revolutionize investing? In their January 2023 paper entitled “How Can Machine Learning Advance Quantitative Asset Management?”, David Blitz, Tobias Hoogteijling, Harald Lohre and Philip Messow study application of machine learning to investing from the perspective of a prudent practitioner. They define machine learning models as initially guided software programs that subsequently learn by themselves and then make predictions about unseen (out-of-sample) data. They describe benefits and pitfalls of machine learning versus classical (linear regression) econometrics. They discuss critical design choices for applying machine learning models to asset management. They focus on the ability of machine learning to beat the stock market, but also discuss asset risk forecasting, optimal portfolio construction and trading optimization. Based on the body of research and their collective experience in asset management, they conclude that: Keep Reading

What is the critical success factor for experienced traders? In their December 2022 paper entitled “Strategic Sophistication and Trading Profits: An Experiment with Professional Traders”, Marco Angrisani, Marco Cipriani and Antonio Guarino compare results from a competitive trading game, a competitive guessing game and individual cognitive/risk preference/personality tests to determine what characteristics most strongly relate to trading success. They recruit two groups for testing: (1) professional traders and portfolio managers working in London financial markets; and, (2) as a control, a gender-matched (mostly male), multi-disciplinary group of undergraduate students with little or no experience trading real markets. For each group, they conduct:

1. A competitive group trading game wherein participants trade an asset that earns dividends over multiple periods in a continuous double auction. Profits derive from both dividends accrued (fundamental value) and any capital gain (ability to outwit other players by buying low and selling high).
2. A competitive group guessing game wherein each participant chooses a number from 0 to 100 with the goal of guessing the number is closest to two thirds of the average choice. Choices reveal how deeply participants reflect on the thinking of others.
3. An individual guessing game (to disentangle reasons for group guessing game outcomes), two tests of cognitive ability (intelligence), tests to elicit risk preferences/confidence and personality traits.

Using game/test results for 56 professional traders and 56 undergraduate students, they find that: Keep Reading

Should investors believe that long-term asset class return forecasts are useful? In his brief August 2022 paper entitled “How Accurate are Capital Market Assumptions, and How Should We Use Them?”, Mike Sebastian employs 10 years of annual Survey of Capital Market Assumptions by Horizon Actuarial Services to assess the industry’s ability to gauge 10-year future asset class returns. This survey presents inputs from leading consulting and investment management firms and includes composite, minimum and maximum forecasted returns for 15 asset classes. Using forecast data for 2012 through 2021, he finds that: Keep Reading

Is past rolling maximum drawdown, a simple measure of recent downside risk, a useful indicator of future mutual fund performance? In their June 2022 paper entitled “Maximum Drawdown as Predictor of Mutual Fund Performance and Flows”, Timothy Riley and Qing Yan investigate whether style-adjusted maximum drawdown based on daily returns over the last 12 months usefully predicts mutual fund performance. To adjust for fund style differences, they subtract from each individual unadjusted drawdown the average unadjusted drawdown across all funds in the same style during the measurement interval. Their principal performance metric is alpha based on a 4-factor (market, size, book-to-market, momentum) model of stock returns. Using daily net returns for 2,188 actively managed long-only U.S. equity mutual funds that are at least two years old and have at least $20 million in assets during January 1999 through December 2019, they find that: Keep Reading

Do presumably expert early-stage startup investors, whether individuals (Angels) or institutions (Venture Capitalists) invest efficiently? In his June 2022 paper entitled “Predictably Bad Investments: Evidence from Venture Capitalists”, Diag Davenport applies machine learning methods based on information known at the time of investment to evaluate decisions of early-stage investors. He defines early-stage investments as equity deals within two years of incubator completion categorized in Pitchbook as deal types Series A, Series B, Seed Round or Angel (Individual). He define late-stage exit as initial public offering, merger/acquisition or funding categorized in Pitchbook as Series C or later. He uses his first five years of quantitative data and numerical transformations of the qualitative data (text) in training a model with XGBoost to predict future venture success. He then applies the model to the next three years of data to build a portfolio that substitutes conventional investments (such as the S&P 500 Index) for predictably bad ventures. Using venture financials and qualitative information about the CEO from Pitchbook for 16,054 startups accepted into top accelerator programs during 2009 through 2016 (2009-2013 for model training and 2014-2016 for testing), he finds that:

Keep Reading

Do individual investors who chase stocks with high recent returns benefit from momentum or suffer from reversal? In their June 2022 paper entitled “Who Chases Returns? Evidence from the Chinese Stock Market”, Weihua Chen, Shushu Liang and Donghui Shi investigate the characteristics, performance and market impact of retail stock investors who exhibit return-chasing behavior. Each month, they measure:

1. Each retail investor’s return chasing propensity (RCP) as the average of returns during the 12 months prior to purchase across the stocks in the investor’s portfolio. For robustness they also consider past return intervals of one, two, three and six months.
2. Each stock’s return chasing ownership (RCO) by wealth-weighting the RCPs of its retail holders (excluding this stock from holder RCP calculations).

Using monthly stock holdings, trading records and investor demographics, plus associated monthly stock prices, for 18 million Shanghai Stock Exchange retail investors during January 2011 through December 2019, they find that:

Keep Reading

Why do the alphas generated by historical simulations/backtests disappear in live trading, with asset managers and brokers the only winners via fees and commissions. In their February 2022 paper entitled “Where’s the Beef?”, Robert Arnott, Amie Ko and Lillian Wu explore: (1) the ways that seasoned professionals fall prey to the simple blunders of data snooping and performance chasing; and, (2) how the industry could actually meet client expectations. Based on the body of research on investor behavior and fund performance and decades of investment management experience, they conclude that: Keep Reading

How have active equity investment managers performed over the past three decades? In his November 2021 paper entitled “Active Equity Management, 1991-2020”, Gene Hochachka examines whether: (1) active equity managers as a group beat their benchmarks over the last 30 years; and, (2) active equity manager relative performance is persistence. By active equity managers, he means:

• Live and dead U.S. mutual funds tracked by Morningstar Direct and not classified as an index fund or fund-of-funds, segmented into US LargeCap, US MidCap, US SmallCap and Foreign (International) LargeCap.
• Institutional strategies tracked as self-reported by Mercer Global Investment Manager Database and not classified as passive in mid-2021, segmented into US LargeCap, US MidCap, US SmallCap, US Small/MidCap, US AllCap and International LargeCap.

Fund/strategy and benchmark returns are for calendar years, including dividends/distributions, and are gross of all fees and expenses. Some analyses compare net-of-expense fund/strategy and net-of-expense benchmark returns. Using the specified annual returns during 1991 through 2020, he finds that: Keep Reading

Does actual performance support the view that university endowments are exemplary stewards of multi-asset class portfolios? In his November 2021 paper entitled “The Modern Endowment Story: A Ubiquitous U.S. Equity Risk Premium”, Richard Ennis re-examines aggregate allocations and performance of U.S. educational endowments. Specifically, he:

• Estimates effective aggregate endowment asset class allocations over different recent sample periods via multiple regressions of endowment returns versus returns of three indexes: Bloomberg Aggregate U.S. bonds; Russell 3000 stocks; and, currency-hedged MSCI ACWI ex-U.S. stocks.
• Applies these effective allocations to construct benchmark portfolios of these three indexes for the different sample periods.

Using investment data for over 100 U.S. educational endowments with assets over $1 billion during the 13 years ending June 2021, he finds that: Keep Reading

What is the best way to conduct quantitative investment research? In his September 2021 presentation package entitled “Escaping The Sisyphean Trap: How Quants Can Achieve Their Full Potential”, Marcos Lopez de Prado outlines the optimal way to tackle such research. Based on his experience, he concludes that: Keep Reading