Giving Color to the Financial Markets: Macroeconomic Information Engineering Through Machine Learning and Portfolio Optimization

Abstract

Trading strategies of financial institutions as well as hedge funds take into account different factors depending on whether investments are in the equity markets or fixed income markets. This research will focus on how different kinds of macroeconomic variables contribute to both equities and fixed income trading. Economic indicators play major roles in influencing the trading of United States fixed income in the interest rate space { such indicators include the unemployment rate, Consumer Price Index (CPI) and gross domestic product (GDP). Central bank policies, foreign exchange prices, and global market indices contribute to the investment strategies of equity macro trading desks. The first part of this research leverages machine learning methods such as Random Forest and Gradient Boosting, in addition to color-coded correlation matrix visualizations incorporating hierarchical clustering algorithms, to better understand the moving parts behind United States economic indicators and global market variables. The second part of this research applies linear programming techniques to generate optimal portfolios comprising domestic and international assets. Moreover, this study considers how the Federal Open Market Committee's interest rate decisions shift the correlations among equities, currencies, commodities, and government bonds.