There are many ways one can lose money when investing, and exchange rates are one of the potential risk factors. Luckily, there are several ways to minimize this type of loss in your portfolio. Systematic FX hedging that uses currency factor strategies is a way of protecting an existing or anticipated position from an unwanted move in an exchange rate. It does not eliminate the risk of loss completely but helps to manage currency exposure better.

Order book information is crucial for traders, but it can be complex. With the numbers of stocks listed in stock exchanges, it is impossible to track all the available information for the human mind. Therefore, the order flows could be an interesting dataset for machine learning models. The novel research of Kolm, Turiel and Westray (2021) utilizes deep-learning for high-frequency return forecasts for 115 NASDAQ stocks based on order book information at the most granular level.

The paper has several key contributions. Firstly, it does not forecast one single return but rather a whole vector of returns – a term structure consisting of mid-price return forecasts at a specified horizon. The forecasted term structure provides essential information about the most optimal execution algorithms (or a trading strategy). According to the authors, forecasts have an „accuracy peak“ at two price changes, after which the accuracy declines. Secondly, the paper compares several methods: autoregressive model with exogenous inputs, MLP, LSTM, LSTM-MLP, stacked LSTM, and CNN-LSTM. Therefore, the article could also serve as a horse race across several possible forecasting methods. Lastly, using more traditional statistical approaches, the authors have identified a better forecasting performance in more information-rich stocks. As a result, this novel research could benefit many areas such as high-frequency trading (but trading costs must be considered), optimal execution strategy, or market-making.

Hello all,

What have we accomplished in the last month?

– A new Portfolio Clustering Quantpedia Pro report
– 13 new Quantpedia Premium strategies have been added to our database
– 10 new related research papers have been included in existing Premium strategies during the last month
– Additionally, we have produced 10 new backtests written in QuantConnect code
– And finally, 5+3 new blog posts that you may find interesting have been published on our Quantpedia blog in the previous month

This is the third and final article from the clustering series. If you’ve missed the previous parts, here you can find the first and second parts of the series. This section examines trading strategies based on previously introduced clustering methods. The complete Portfolio Clustering report will be available for our Quantpedia Pro clients next week.

The CAPM model was a breakthrough for asset pricing, but the times where the market factor was most widely used are long gone. Nowadays, if we exaggerate a bit, we have as many factors as we want. Therefore, it might not be straightforward which factor model should be used. 

Hanauer et al. (2021) provide several insights into factor models. The authors postulate that the factor models should be examined in the international samples since this can be understood as a test for asset pricing models. The domestic Chinese A-shares stock market seems to be an excellent “playground” for the factors models, given the size of the Chinese stock market, but mainly because of its uniqueness. The paper compares the models (and factors) based on various methods (performance, data-driven asset pricing framework, test assets, turnovers and even transaction costs). Apart from valuable insights into the several less-known factors, the key takeaway message could be that the “US classic” Fama-French factor models perform poorly in China. The modified Fama-French six-factor model or q-factor is better, but overall, it seems that factor models designed for China, such as the model of Liu, Stambaugh and Yuan (2019), are the best.

October’s is coming, and we continue our short series of introductory articles about portfolio clustering methods we will soon use in our new Quantpedia Pro report. In the previous blog, we introduced three clustering methods and discussed the pros and cons of each one. Additionally, we showed a few examples of clustering, and we presented various methods for picking an optimal number of clusters.

This section demonstrates the Partitioning Around Medoids (PAM) – a centroid-based clustering method, Hierarchical Clustering, which uses machine learning and Gaussian Mixture Model based on probability distribution and applies all three methods to an investment portfolio that consists of eight liquid ETFs.