The first part of this study began from an intuitive proposition: financial markets are networks, and models capable of learning from those networks should therefore possess an informational advantage over models that treat securities independently.
This article is the first in Trading the Network, a seven-part series investigating whether relationships across financial assets contain predictive information that can ultimately be converted into tradable signals. Part I establishes the baseline by asking whether a simple network of historical return correlations
Download PDF Introduction As announced in our previous episode, the second article in our Backtesting series will discuss Cross Validation and introduce a new Backtesting strategy, first presented by Giuseppe Paleologo (2024). If you haven’t already, we recommend reading the first episode before proceeding with this article. Please note that Read more…
Download PDF Introduction When managing a portfolio, the fundamental problem is to find investments with excess risk-adjusted expected rates of return. Once the manager has identified some favourable opportunities and is ready to assemble the portfolio, they still need to decide how much capital, as a fraction of the value Read more…
Download PDF Introduction In this article, we discuss an equity Pairs Trading strategy using the S&P 500 index components. Pairs Trading is a statistical arbitrage strategy that involves simultaneously taking long and short positions in two financial assets, expecting their spread to converge to the mean. However, the process of Read more…
Differences between historical realized volatility and option implied volatility has for long been known as a potential indicator for volatility mispricing. To our knowledge, the first well-known academic study on the topic was conducted by Goyal and Saretto in 2009 and published as the paper ‘Cross-section of Option Returns and Read more…
In finance, a covariance matrix can be a useful tool to estimate the cross correlations between different stocks. However, historical data contains random noise, which can alter the underlying information. Here we illustrate how Random Matrix Theory can be used to filter a diagonalisable matrix and how information about the market Read more…