Quantpedia Premium Update – 18th April

New strategies:

#859 – Earnings Delays and Advances Predict Stock Returns

Period of rebalancing: Daily
Markets traded: equities
Instruments used for trading: stocks
Complexity: Complex strategy
Backtest period: 2006-2020
Indicative performance: 8%
Estimated volatility: 6.68%

Source paper:

Hafez, Peter and Matas Navarro, Ricard and Grinis, Inna and Gomez, Francisco: Trading Around the Earnings Calendar
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4323541
Abstract:
Corporate executives and investor relations teams often use strategic timing to schedule and disseminate earnings results. Studies have shown that earnings delays may signal weak performance, while advancing the date may be a sign of good news. In this paper, we study the insights contained in the date changes of earnings releases, tracked using the RavenPack Earnings Dates dataset. (1) We show that advances/delays in earnings announcement dates can be predictive of positive/negative earnings results. (2) We also show that changes in earnings announcement dates are followed by outsized price reactions, with the resulting strategy achieving robust returns with a slow decay. (3) A combined strategy of earnings announcement events and earnings calendar change events delivers Annualized Returns of 7.9% for US Mid/Large-Caps and 20.7% for Small-Caps, with Information Ratios of 0.8 and 1.4, respectively. (4) Including news sentiment as an overlay, we were able to further enhance strategy performance, bringing Annualized Returns to 8.4% for Mid/Large and 24.6% for Small-Caps.

#860 – Machine learning Analysts’ Sentiment Industry Factor

Period of rebalancing: Monthly
Markets traded: equities
Instruments used for trading: ETFs, stocks
Complexity: Very complex strategy
Backtest period: 1997-2021
Indicative performance: 8.21%
Estimated volatility: 15.97%

Source paper:

Chhaochharia, Vidhi and Kumar, Alok and Rantala, Ville and Zhang, Alan: Artificially Intelligent Analyst Sentiment and Aggregate Market Behavior
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4249442
Abstract:
This study develops a new machine learning-based measure of aggregate analyst sentiment. We first train analyst-specific neural network models that capture each analyst’s predictable forecast bias across firms and then aggregate the information at the industry and market levels. We decompose the aggregated forecast errors of analysts into predictable and non-predictable components, and interpret the non-predictable component as a measure of analyst sentiment. Variation in analyst sentiment along the business cycle suggests that they systematically underreact to macroeconomic information. A Long-Short trading strategy based on industry-level analyst sentiment earns an annualized alpha of over 7%.

#861 – Bond ETF arbitrage with Machine Learning

Period of rebalancing: Daily
Markets traded: bonds
Instruments used for trading: ETFs
Complexity: Very complex strategy
Backtest period: 2002-2020
Indicative performance: 17.64%
Estimated volatility: 5.31%

Source paper:

Crego, A. Julio and Kvaerner, Jens Soerlie and Sommervoll, Åvald and Sommervoll, Dag Einar and Stevens, Niek: Evolutionary Arbitrage
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4051930
Abstract:
The prices of exchange-traded funds (ETFs) can deviate significantly from their net asset values (NAVs). Exploiting such inefficiencies is often too costly because it involves taking positions in hundreds of underlying illiquid securities. We develop a method that identifies a liquid mimicking portfolio that tracks the NAV using only ETFs. Our method combines a genetic algorithm with non-negative least squares. We apply it to the fixed income ETF market. Our long-short strategy generates a Sharpe ratio of 4-5, incurs little transaction cost, and does well under all market conditions.

#862 – Switching Between Momentum and Reversal Strategies Based on Market Volatility

Period of rebalancing: Monthly
Markets traded: equities
Instruments used for trading: stocks
Complexity: Complex strategy
Backtest period: 1932-2020
Indicative performance: 19%
Estimated volatility: 20.91%

Source paper:

Butt, Hilal Anwar and Kolari, James W. and Sadaqat, Mohsin: Momentum, Market Volatility, and Reversal
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4342008
Abstract:
Momentum profits collapse and reversal occurs when preceding market volatility is relatively high. Based on these intertemporal patterns, we implement an investment strategy that switches from momentum to reversal when volatility is high. The proposed switching strategy has two advantages over scaled momentum strategies: (1) the leverage factor is constant, and (2) no ex-post information is used to control for volatility. In U.S. stock market tests across a variety of performance metrics, the switching strategy distinguishes itself from traditional and volatility scaled momentum strategies by eliminating losses due to momentum crashes. Further evidence confirms that the switching strategy is successful in other developed and emerging stock markets, especially in Japanese and Chinese stock markets.

#863 – Persistence of Abnormal Trading Volume Effect

Period of rebalancing: Monthly
Markets traded: equities
Instruments used for trading: stocks
Complexity: Complex strategy
Backtest period: 1970-2020
Indicative performance: 20.27%
Estimated volatility: 15.69%

Source paper:

Li, Mingyi and Yin, Xiangkang and Zhao, Jing: Persistence or Reversal? the Effects of Abnormal Trading Volume on Stock Returns
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4346340
Abstract:
After documenting that monthly portfolios constructed according to the extreme deciles of Abnormal Trading Volume (ATV) generate positive (negative) returns in the short (long) run, we introduce a measure of ATV persistence and offer an explanation for this return predictability based on investor sentiment. ATV persistence leads portfolio returns to keep drifting in the short run. However, the ATV of individual stocks in the portfolio gradually reverts to the long-run mean, accompanied by portfolio returns falling and turning negative as mispricing is corrected. We also invalidate the theories of liquidity shocks and continuing overreaction in explaining the observed return predictability.

New research papers related to existing strategies:

#85 – Momentum in Mutual Fund Returns

Cuthbertson, Keith and Nitzsche, Dirk and O’Sullivan, Niall: UK Mutual Funds: Performance Persistence and Portfolio Size
https://ssrn.com/abstract=4335266
Abstract:
We re-examine performance persistence amongst UK mutual funds. Specifically, we investigate performance persistence among small portfolios of past high-performing funds. In contrast to the more common analysis of decile portfolios of funds, we focus on persistence in the more extreme positive tail of the cross-section of fund performance. This paper contributes to the smaller literature on UK rather than US mutual fund performance. We investigate fund persistence based on practitioner index models as well as academic factor models, focusing on small portfolios of funds using inference based on nonparametric persistence test statistics as well as conventional t-tests. We provide strong evidence of positive persistence among small-size portfolios of (past) high performing funds that is robust to alternative formation and holding periods and alternative performance models. We also document some sensitivity in inferences on positive persistence when using nonparametric versus conventional tests.

#823 – Machine Learning and the Cross-Section of Cryptocurrency Returns

Liu, Jing and Kang, Yuncheol: Automated Cryptocurrency Trading Approach Using Ensemble Deep Reinforcement Learning: Learn to Understand Candlesticks
https://ssrn.com/abstract=4348791
Abstract:
Despite its high risk, cryptocurrency has gained popularity as a successful trading option. Cryptocurrencies are digital assets that fluctuate dramatically in a market that operates for 24 hours. Developing trading bots using machine learning-based artificial intelligence (AI) approaches has recently received considerable attention. Previous studies have used machine learning techniques to predict financial market trends or make trading decisions, primarily using numeric data extracted from candlesticks. However, these data often ignore the temporal and spatial information present in candlesticks, resulting in a poor understanding of their significance. In this study, we used multi-resolution candlestick images that contain temporal and spatial information on prices. The goal of this study was to compare the performance of raw numeric data and candlestick image data to optimize trading strategies and maximize returns. We used deep reinforcement learning algorithms (Deep Q-Networks (DQN), Dueling-DQN, and Proximal Policy Optimization (PPO)) to generate trading signals for opening a long or short position, closing a position, or staying idle. The trading signal was generated using a multiagent weighted voting ensemble approach. We tested the ensemble automated trading approach on two BTC/USDT datasets, a 30-day bullish market, and a 15-day bearish market. Our findings showed that models using candlestick image data outperformed those using numeric data and other baseline models. Additionally, we used a visual representation of candlestick images to depict the results from attention-based deep reinforcement learning algorithms, highlighting its advantages over other models.

#658 – Betting Against Uncertainty Beta in Australia
#676 – Short-Term Reversal and High Uncertainty Periods

Jiang, Fuwei and Kang, Jie and Meng, Lingchao: Certainty of Uncertainty for Asset Pricing
https://ssrn.com/abstract=4390871
Abstract:
Uncertainty is known to be crucial in asset pricing, yet evidence from comprehensive analysis of various uncertainty measures remains sparse. This paper investigates the predictability of stock returns based on economic fundamentals uncertainty by constructing a novel uncertainty index derived from a heterogeneous range of uncertainty measures utilizing a machine learning method. Our composite uncertainty index exhibits robust in- and out-of-sample predictability of stock market returns over the one- to 12-month horizon. The predictive power stems from the individual uncertainty measures significantly correlated with market volatility, and becomes more pronounced during high uncertainty and recession periods. The predictability of our uncertainty index aligns with theoretical frameworks linking uncertainty to future investment, cash flows, and market expectations.

#827 – A Machine Learning Approach to Stock Returns Prediction in China

Ma, Tian and Wang, Wanwan and Chen, Yu: Attention Is All Your Need. the Evidence from Asset Allocation with Interpretable Transformer Model
https://ssrn.com/abstract=4329294
Abstract:
Deep learning technology is rapidly adopted in financial market settings. Using a large data set from the Chinese stock market, we propose a return-risk trade-off strategy via transformer model. The empirical finding shows these updates in technology can improve the use of information related to return and volatility, increase the predictability, and then capture more economic gains than other non-linear models such as LSTM. Our model gives rise to a new measure of “economic feature importance” with SHAP value and tabulates the different important features between return and risk prediction. Last, we provide several economic explanations for robust. This paper sheds lights on the burgeoning field on asset allocation in the age of Big Data.

#546 – Implied Volatility Spreads and Expected Market Returns in S&P500
#710 – Quantile Curves and the VRP

Freire, Gustavo and Kleen, Onno: Equity Options and Firm Characteristics
https://ssrn.com/abstract=4342597
Abstract:
We study the relation between a comprehensive set of firm characteristics and the entire universe of individual equity option prices. We find that 42 out of 86 characteristics are priced in the option market, in the sense that they significantly explain differences in the implied volatility surface (IVS) across stocks. Motivated by this finding, we model the IVS of a given stock as a function of its characteristics with a local linear random forest. This approach addresses the illiquidity of the equity option market by effectively grouping similar stocks during estimation. Our method outperforms a stock-specific benchmark model out-of-sample and allows us to uncover the nonlinear interactions between characteristics and option prices.

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

BERT Model – Bidirectional Encoder Representations from Transformers

At the end of 2018, researchers at Google AI Language made a significant breakthrough in the Deep Learning community. The new technique for Natural Language Processing (NLP) called BERT (Bidirectional Encoder Representations from Transformers) was open-sourced. An incredible performance of the BERT algorithm is very impressive. BERT is probably going to be around for a long time. Therefore, it is useful to go through the basics of this remarkable part of the Deep Learning algorithm family.

Are Funds Flows Influenced by Mortality?

Countries from the majority of the developed world face one challenge: Their population is steadily aging; the average age of individuals has been rising over recent periods. The United States is not different in this sense. Whenever there’s a never-inevitable reaching of higher ages, people reconsider their choices and often cut on riskier ones. So, is there a potential link between demographic changes associated with aging and aggregate financial market outcomes?

Political Beliefs Matter for Fund Managers

Two leading political parties, the Democrats and Republicans, have dominated the United States politic for decades. As a consequence, the significant differences in views on major issues of partisans from different parties may influence their economic expectations. Recent studies found that partisan politics significantly impacts household beliefs and economic decision-making. But do political beliefs matter to institutional investors?