S&P500 Forecasting and Trading using Convolution Analysis of Major Asset Classes

📝 Abstract

By monitoring the time evolution of the most liquid Futures contracts traded globally as acquired using the Bloomberg API from 03 January 2000 until 15 December 2014 we were able to forecast the S&P 500 index beating the Buy and Hold trading strategy. Our approach is based on convolution computations of 42 of the most liquid Futures contracts of four basic financial asset classes, namely, equities, bonds, commodities and foreign exchange. These key assets were selected on the basis of the global GDP ranking across countries worldwide according to the lists published by the International Monetary Fund (IMF, Report for Selected Country Groups and Subjects, 2015). The main hypothesis is that the shifts between the asset classes are smooth and are shaped by slow dynamics as trading decisions are shaped by several constraints associated with the portfolios allocation, as well as rules restrictions imposed by state financial authorities. This hypothesis is grounded on recent research based on the added value generated by diversification targets of market participants specialized on active asset management, who try to efficiently and smoothly navigate the market’s volatility.

💡 Analysis

By monitoring the time evolution of the most liquid Futures contracts traded globally as acquired using the Bloomberg API from 03 January 2000 until 15 December 2014 we were able to forecast the S&P 500 index beating the Buy and Hold trading strategy. Our approach is based on convolution computations of 42 of the most liquid Futures contracts of four basic financial asset classes, namely, equities, bonds, commodities and foreign exchange. These key assets were selected on the basis of the global GDP ranking across countries worldwide according to the lists published by the International Monetary Fund (IMF, Report for Selected Country Groups and Subjects, 2015). The main hypothesis is that the shifts between the asset classes are smooth and are shaped by slow dynamics as trading decisions are shaped by several constraints associated with the portfolios allocation, as well as rules restrictions imposed by state financial authorities. This hypothesis is grounded on recent research based on the added value generated by diversification targets of market participants specialized on active asset management, who try to efficiently and smoothly navigate the market’s volatility.

📄 Content

S&P500 FORECASTING AND TRADING USING CONVOLUTION ANALYSIS OF MAJOR ASSET CLASSES

Panagiotis Papaioannou1,Thomas Dionysopoulos2, Dietmar Janetzko3, Constantinos Siettos1,*

1School of Applied Mathematics and Physical Sciences, National Technical University of Athens, Greece *Corresponding author 2AXIANTA RESEARCH, Nicosia, Cyprus - Avenir Finance Investment Managers, Paris, France 3Cologne Business School, Cologne, Germany

Abstract

By monitoring the time evolution of the most liquid Futures contracts traded globally as acquired using the Bloomberg API from 03 January 2000 until 15 December 2014
we were able to forecast the S&P 500 index beating the “Buy and Hold” trading strategy. Our approach is based on convolution computations of 42 of the most liquid Futures contracts of four basic financial asset classes, namely, equities, bonds, commodities and foreign exchange. These key assets were selected on the basis of the global GDP ranking across countries worldwide according to the lists published by the International Monetary Fund (IMF, Report for Selected Country Groups and Subjects, 2015). The main hypothesis is that the shifts between the asset classes are smooth and are shaped by slow dynamics as trading decisions are shaped by several constraints associated with the portfolios allocation, as well as rules restrictions imposed by state financial authorities. This hypothesis is grounded on recent research based on the added value generated by diversification targets of market participants specialized on active asset management, who try to efficiently and “smoothly navigate” the market’s volatility.

Keywords: Financial Assets; Forecasting; Quantitative Trading Strategies;Financial Time Series

1. INTRODUCTION

Corporate finance theorists, macroeconomists, behavioral psychologists, quantitative finance mathematicians have teamed up to find the holy grail of the financial markets: that of consistently beating the market. Towards this aim, methods can be categorized into (a) fundamentals, and (b) financial engineering. On the one hand, Fundamentals include models from corporate finance theories, (e.g. Capital Asset Pricing Model: Fama and French, 1989) and macroeconomic theories (e.g. the Keynesian model: Keynes, 1936; Bodkin and Eckstein, 1985 etc.) most of which try to approximate the dynamics of macroeconomic observables such as the aggregate demand and supply, investment volume and consumption, risk premia etc. On the other hand, financial engineering and quantitative finance models, that have been flourished after the seminal work of the derivatives contracts pricing model (Black and Scholes, 1973) try to forecast the price action of several financial instruments such as the S&P 500 Index (Niaki and Hoseinzade, 2013) on a data driven basis. Both approaches focus basically on several sub-targets such as the forecasting of a single asset as well as asset allocation methods among the four basic asset classes, namely, Equities, Commodities, Bonds, and Foreign Exchange Markets (Bekkers, et al., 2012). Several models have been proposed to forecast their future dynamics based on historical data and information acquired from almost all possible sources: price action, fundamentals as well as behavioral sentiment analysis and e-social platforms (Husain and Bowman, 2004; Polk, et al., 2006; Diebold and Li, 2006; Papaioannou et al., 2013). Even though success stories have been reported, these models are still prone to failure in situations undergoing structural changes and market crashes (Zellner&Chetty, 1965; Klein &Bawa, 1976; Kandel&Stambaugh, 1996; Barberis, 2000; Caballero et al., 2008). Here we propose an approach to forecast the E-mini S&P500 Futures contract exploiting convolution analysis of 42 of the most liquid Futures contracts of four basic financial asset classes: (1) equities, (2) bonds, (3) commodities, and, (4) foreign exchange.Our choice is motivated by the fact that we want to use a “non-memory” approach, one that does not need training datasets (e.g. ANNs etc.) and which would provide a pointwise, time dependent and out-of-sample trading strategy even from the first few observations of the historical dataset. Based on the proposed approach, we managed to successfully forecast the S&P500 Futures Contract beating the “Buy and Hold” benchmark trading strategy.

2. METHODOLOGY

2.1 The Hypothesis

The question of how financial uncertainty gets incorporated in the risk premia offered by several financial assets and how it formulates the investors’ preferences and their corresponding portfolios allocation is fundamental in contemporary financial research. Previous studies have shown that allocation decisions made by fund managers on behalf of the investors are shaped by several bounds regarding the portfolios weights of each asset class in order to ensure portfolio robustness, lower transaction costs (turnovers

This content is AI-processed based on ArXiv data.