Out-of-sample Dataset Before the “Sample”: Pervasive Anomalies Before 1926

30.November 2021

Data are the key to systematic investing/trading strategies. The hypotheses testing, risk or return evaluations, correlations, and factor loadings rely on past data and backtests. With an increasing speed of publication in finance, critiques of quantitative strategies have emerged. Strategies seem to decay in alpha, post-publication returns tend to be lower, and many strategies become insignificant once rigorously tested (in or out-of-sample). Moreover, some might even appear profitable purely by chance and the repetitive examination of the same dataset, such as CRSP stocks after 1963. 

Is there any solution to overcome these limitations? Partially, the design of the novel machine learning strategies consisting of training, validation, and testing sets might help. Perhaps the most crucial part of such a scheme is the usage of the purely out-of-sample dataset. In this regard, the novel research by Baltussen et al. (2021) provides several valuable findings for the most recognized factors. The authors constructed a database of U.S. stocks, including dividends and market caps for 1488 major stocks from 1866 to 1926. The sample can be described as the pre-CRSP period, including independent, pre-publication, and “out-of-sample” data that can be a perfect test for the factors utilized today. 

Continue reading

The Quant Cycle – The Time Variation in Factor Returns

22.November 2021

Although the factors in asset pricing models offer a premium in the long run, they are undergoing bull and bear market cycles in the short term. One would expect that it is due to their connection to the business cycles as the factor premium represents a reward for bearing the macroeconomic risks. A novel study by Blitz (2021) finds that traditional business cycle indicators can’t explain much of the time variation of factor returns as the factors are a behavioral phenomenon driven by investor sentiment. To capture the large factor cyclical variation, the author proposes a quant cycle that is defined by the peaks and troughs in the factor returns corresponding to the bull and bear markets.

Continue reading

Community Alpha of QuantConnect – Part 4: Composite Social Trading Multi-Factor Strategy

18.November 2021

This blog post is the continuation (and finale) of series about Quantconnect’s Alpha market strategies. This part is related to the multi-factor strategies notoriously known from the majority of asset classes. We continue in the examination of factor strategies built on top of social trading strategies, but the investment universe is reduced based on the insights of the previous part. So, without further ado, we continue where we have left last time.

Continue reading

How to Combine Different Momentum Strategies

15.November 2021

Today we will again talk more about the portfolio management theory, and we will focus on techniques for combining quantitative strategies into one multi-strategy portfolio. So, let’s imagine we already have a set of profitable investment strategies, and we need to combine them. The goal of such “strategy allocation” usually is to achieve the best risk-adjusted return possible. There is no single correct solution to this task, but there are a few methods that we can try.

The “appropriate combination” highly depends on the type of strategies we are about to combine. Are we combining equity and bond strategies together? Are we combining equity strategies, with each one having an entirely different logic? Or do we rather need to assign weights to strategies that are similar in nature yet still different? We will focus this article on the last option – combining similar yet different strategies.

Continue reading

How News Move Markets?

12.November 2021

Nobody would argue that nowadays, we live in an information-rich society – the amount of available information (data) is constantly rising, and news is becoming more accessible and frequent. It is indisputable that this evolvement has also affected financial markets. Machine learning algorithms can chew up big chunks of data. We can analyze the sentiment (which is frequently related to the news). Big data does not seem to be a problem anymore, and high-frequent trading algorithms can react almost instantly. But how important is the news? Kerssenfischer and Schmeling (2021) provide several answers by studying the impact of scheduled and unscheduled news (frequently omitted in other news-related studies) in connection with high-frequency changes in bond yields and stock prices in the EU and US as well. The research points out that the effect is tremendous and significant.

Continue reading

Bitcoin Returns and Volatility Predicted by Bitcoin Exchange Reserves

9.November 2021

In the modern world full of technologies, cryptocurrencies are gaining popularity every day. The most famous cryptocurrency, bitcoin, was introduced in 2009. Ever since its launch and its subsequent success, when within a few years, its price skyrocketed, and it has been the subject of many price predicting studies. These, however, primarily focus on the market and macro factors, entirely omitting the nature of bitcoin – which is blockchain technology. In this study, authors Hoang and Baur try to capture and research this interconnection between behaviour of investors, bitcoin exchanges, and blockchain.

Continue reading

Subscribe for Newsletter

Be first to know, when we publish new content


    logo
    The Encyclopedia of Quantitative Trading Strategies

    Log in