Quantpedia Premium Update – August 23rd

New Strategies:

#1038 – Presidential Cycle and Performance of the Stock Market

Period of rebalancing: 3 years
Markets traded: equities
Instruments used for trading: CFDs, ETFs, funds, futures, stocks
Complexity: Simple strategy
Backtest period: 1926-2020
Indicative performance: 12.33%
Estimated volatility: 8.97%

Source paper:

Rebecca Osamudiame, Oumar Sy: Presidential-Partisan Cycles and the Cross-Section of Stock Returns
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4876493
Abstract:
Abstract: from 1926 to 2020, average excess stock return differentials between Democratic and Republican presidencies were positive for more than two-thirds of publicly traded US firms, averaging about 12% per annum across all firms. Democratic presidents outperform their Republican counterparts primarily for unprofitable, illiquid, inefficient, distressed, and financially constrained small and medium-sized enterprises. Democratic presidencies notably more favorably impact firms in the oil and telecommunications industries, while gun-related firms thrive under Republican presidencies. Presidential politics appears less relevant to large corporations or information asymmetries mitigation and to how investors value capital assets or how firms make their investment and financing decisions.

#1039 – Euro Momentum Strategy

Period of rebalancing: Monthly
Markets traded: equities
Instruments used for trading: CFDs, ETFs, futures
Complexity: Moderately complex strategy
Backtest period: 2000-2020
Indicative performance: 9.99%
Estimated volatility: 13.12%

Source paper:

Vukovic; Darko B.; Ingenito; Salvatore; Maiti; Moinak: Time series momentum: Evidence from the European equity market
https://doi.org/10.1016/j.heliyon.2023.e12989
Abstract:
This study empirically analyzes time series momentum (TSM) in the European equity market between 2000 & 2020. The study produces additional evidence on TSM where a significant and persistent market price anomaly enables investors to earn abnormal returns. To achieve this goal the present study implements a pooled autoregressive model to test the predictability power of European equity indices of future returns. The results indicate that strategies based on TSM are in line with the discussed literature and enable market agents to earn returns above the market (0.71% per month) by using a six-factor model.

#1040 – Price-to-Utility Ratio in Bitcoin

Period of rebalancing: Daily
Markets traded: cryptos
Instruments used for trading: cryptos
Complexity: Very complex strategy
Backtest period: 2013-2022
Indicative performance: 55.76%
Estimated volatility: 14.95%

Source paper:

Yu; Haoyang; Sun; Yutong; Liu; Yulin; Zhang; Luyao: Bitcoin Gold, Litecoin Silver: An Introduction to Cryptocurrency’s Valuation and Trading Strategy
https://www.semanticscholar.org/reader/fc2c0d4f6483753851d7653b8fea682ab31f8af8
Abstract:
Historically, gold and silver have played distinct roles in traditional monetary systems. While gold has primarily been revered as a superior store of value, prompting individuals to hoard it, silver has commonly been used as a medium of exchange. As the financial world evolves, the emergence of cryptocurrencies has introduced a new paradigm of value and exchange. However, the store-of-value characteristic of these digital assets remains largely uncharted. Charlie Lee, the founder of Litecoin, once likened Bitcoin to gold and Litecoin to silver. To validate this analogy, our study employs several metrics, including unspent transaction outputs (UTXO), spent transaction outputs (STXO), Weighted Average Lifespan (WAL), CoinDaysDestroyed (CDD), and public on-chain transaction data. Furthermore, we’ve devised trading strategies centered around the Price-to-Utility (PU) ratio, offering a fresh perspective on crypto-asset valuation beyond traditional utilities. Our back-testing results not only display trading indicators for both Bitcoin and Litecoin but also substantiate Lee’s metaphor, underscoring Bitcoin’s superior store-of-value proposition relative to Litecoin. We anticipate that our findings will drive further exploration into the valuation of crypto assets. For enhanced transparency and to promote future research, we’ve made our datasets available on Harvard Dataverse and shared our Python code on GitHub as open source.

#1041 – Macroeconomic Announcement Days and Machine Learning

Period of rebalancing: Daily
Markets traded: equities
Instruments used for trading: stocks
Complexity: Very complex strategy
Backtest period: 2005-2020
Indicative performance: 17.84%
Estimated volatility: 19.84%

Source paper:

Liao, Cunfei and Ma, Tian and Jiang, Fuwei: Macroeconomic Announcement and Machine Learning for Asset Pricing
https://ssrn.com/abstract=4824740
Abstract:
Machine learning techniques have significantly improved the prediction of expected stock returns. By distinguishing between trading days with major macroeconomic announcements (A-days) and regular trading days (N-days), we find that machine learning effectively captures time-varying sources of predictability of returns on A-days and N-days. We construct an ensemble model that combines results from models trained separately on these distinct types of days. Notably, the ensemble method outperforms models using complete datasets or subsets in forecasting accuracy, which highlights the “complexity in time-series variation”. Evidence from bond-related characteristics reinforces the presence of time variation in asset pricing models.

New research papers related to existing strategies:

#460 – ESG Level Factor Investing Strategy

Tan, Yeng-May and Szulczyk, Kenneth R and Sii, Yew-Hei: Performance of ESG-Integrated Smart Beta Strategies in Asia-Pacific Stock Markets
https://ssrn.com/abstract=4781519
Abstract:
Environmental, Social, and Governance (ESG) investing is about ethical investing. While ESG investing has garnered heightened attention, the research has not settled on whether ESG investing can “do well while doing good”. Using a proprietary ESG rating database of monthly firm-specific data, we examine the performance of ESG-incorporated investing strategies in Australia, Mainland China, Hong Kong, Malaysia, and Singapore. Specifically, we combine positive screening and the smart beta approach to evaluate the performance of ESG-based and non-ESG-based (traditional) equity portfolios. Our key findings reveal that high-ESG-based portfolios do not offer superior risk-adjusted returns compared to the low-rated portfolios. While a high-ESG-rated portfolio generally outperforms the market index, the ESG and traditional smart beta alphas differ little. Our results also indicate that the minimum-volatility portfolio achieves the best performance of all the factors. We use data from Refinitiv ESG and Bloomberg ESG to substantiate and support the results. Our findings add to the growing ESG literature that answers whether investors risk sacrificing returns while investing ethically.

#874 – Commodity Pairs Trading in China using Machine Learning

Yang, Shuo and Huang, Ke: Research on Hierarchical Futures Pair Trading Strategy Based on Machine Learning and Kalman Filtering
https://ssrn.com/abstract=4863414
Abstract:
Pair trading, as a significant arbitrage strategy, profits from the residuals between asset pairs. This study aims to explore the application of machine learning and Kalman filtering techniques in pair trading to optimize traditional strategies, enhancing both the return and stability of the strategy. Specifically, we address two issues: (1) how to improve the pairing efficiency and stability of assets, and (2) how to reduce noise interference and generate dynamic trading signals that enhance returns. Empirical results demonstrate that the pair trading strategy based on machine learning and Kalman filtering holds a significant advantage in hedge trading. This strategy enhances the cointegration pairing efficiency using the DTW clustering algorithm and employs Kalman filtering to eliminate noise, successfully increasing the accuracy and stability of trading signals. Additionally, utilizing the half-life as a sliding window to calculate the z-value (trading signal) allows the strategy to more flexibly seize trading opportunities. The empirical results indicate that this strategy has achieved favorable returns in backtesting, with robustness and effectiveness that can provide investors with valuable strategic optimization insights.

#700 – Expected Options Return Predictability Using Machine Learning

Höfler, Philipp: Volatility Surfaces and Expected Option Returns
https://ssrn.com/abstract=4869272
Abstract:
This paper applies deep learning techniques to uncover novel return predictability in the cross-section of delta-hedged equity options. I demonstrate sizable profits in long-short option portfolios using a Convolutional Neural Network (CNN) that automatically extracts relevant patterns from the implied volatility surface. Portfolio returns remain statistically and economically significant even after accounting for transaction costs. The CNN subsumes some commonly used predictors but cannot be fully explained by option-based characteristics. I further show that the CNN generates abnormal returns relative to a latent factor model that is based on a wide range of option and stock characteristics. Finally, I provide evidence that the model can also be used to predict returns of alternative option positions such as straddles.

#628 – Social Media Sentiment Factor

Hosseini, Amin and Jostova, Gergana and Philipov, Alexander and Savickas, Robert: The Social Media Risk Premium
https://ssrn.com/abstract=4854532
Abstract:
We show that social media risk is priced in the cross sections of stocks and bonds. New social media stock and bond factors earn annual premiums of 7.2% and 3.3%, respectively. Their contributions to explaining the cross-section are significant when tested both with classical and recent machine-learning asset-pricing methodologies. The social media risk premium is higher when market uncertainty is higher and sentiment is lower. Unlike other risk factors, the social media factor origins are clearly identified: prior to the age of social media their premiums did not exist.

#1028 – Sentiment Analysis of Bloomberg Markets Wrap Using ChatGPT

Lefort, Baptiste and Benhamou, Eric and Ohana, Jean-Jacques and Saltiel, David and Guez, Beatrice and Jacquot, Thomas: Mixing Financial Stress with GPT-4 News Sentiment Analysis for Optimal Risk-On/Risk-Off Decisions
https://ssrn.com/abstract=4781752
Abstract:
This paper introduces a new risk-on risk-off strategy for the stock market, which combines a financial stress indicator with a sentiment analysis done by ChatGPT reading and interpreting Bloomberg daily market summaries. Forecasts of market stress derived from volatility and credit spreads are enhanced when combined with the financial news sentiment derived from GPT4. As a result, the strategy shows improved performance, evidenced by higher Sharpe ratio and reduced maximum drawdowns. The improved performance is consistent across the NASDAQ, the S&P 500 and the six major equity markets, indicating that the method generalises across equities markets.

#906 – Using ChatGPT to Forecast Stock Price Movements

Mai, Dat, StockGPT: A GenAI Model for Stock Prediction and Trading
https://ssrn.com/abstract=4787199
Abstract:
This paper introduces StockGPT, an autoregressive “number” model trained and tested on 70 million daily U.S. stock returns over nearly 100 years. Treating each return series as a sequence of tokens, StockGPT automatically learns the hidden patterns predictive of future returns via its attention mechanism. On a held-out test sample from 2001 to 2023, a daily rebalanced long-short portfolio formed from StockGPT predictions earns an annual return of 119% with a Sharpe ratio of 6.5. The StockGPT-based portfolio completely spans momentum and long-/short-term reversals, eliminating the need for manually crafted price-based strategies, and also encompasses most leading stock market factors. This highlights the immense promise of generative AI in surpassing human in making complex financial investment decisions.

And several interesting free blog posts that have been published during the last 2 weeks:

Payout-Adjusted CAPE

Professor Robert Shiller’s CAPE (cyclically adjusted price-to-earnings) ratio is well-known among the investment community. His methodology for assessing a valuation of the U.S. equity market is undoubtedly the most cited and discussed. Therefore, it’s not surprising that there exists quite a lot of papers that try to refine and expand the CAPE’s methodology. One such last attempt is the work of James White and Victor Haghani, whose research paper revolves around the use of a modified version of the Cyclically-Adjusted Price Earnings (CAPE) ratio, termed P-CAPE. Their methodology aims to improve the estimation of long-term expected real returns of the stock market by incorporating the dividend payout ratio into the traditional CAPE metric.

Lunch Effect in the U.S. Stock Market Indices

In the complex world of financial markets, subtle patterns often reveal themselves through careful observation and analysis. Among these is the intriguing phenomenon we can call the “Lunch Effect,” a pattern observed in U.S. stock indexes where market performance tends to exhibit a distinct positive shift immediately after the lunch break, following a typically negative or flat performance earlier in the trading day right before the lunch. This lunchtime revival is not an isolated occurrence; it shares a curious connection with the “Overnight Effect,” a well-documented tendency for the U.S. stock market to experience the bulk of its appreciation during non-trading hours, with relatively little movement during the trading day itself. Together, these effects underscore the intricate dynamics of market behavior, where timing and investor psychology play crucial roles in shaping intraday and overnight market performance. Understanding these patterns can offer valuable insights into the rhythm of the markets and the underlying factors that drive short-term price movements.