Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 29 July 2021 Article A Linear Process Approach to Short-term Trading Using the VIX Index as a Sentiment Indicator Yawo Mamoua Kobara 1,‡ , Cemre Pehlivanoglu 2,‡* and Okechukwu Joshua Okigbo 3,‡ 1 Western University;
[email protected] 2 Cidel Financial Services;
[email protected] 3 WorldQuant University;
[email protected] * Correspondence:
[email protected] ‡ These authors contributed equally to this work. 1 Abstract: One of the key challenges of stock trading is the stock prices follow a random walk 2 process, which is a special case of a stochastic process, and are highly sensitive to new information. 3 A random walk process is difficult to predict in the short-term. Many linear process models that 4 are being used to predict financial time series are structural models that provide an important 5 decision boundary, albeit not adequately considering the correlation or causal effect of market 6 sentiment on stock prices. This research seeks to increase the predictive capability of linear process 7 models using the SPDR S&P 500 ETF (SPY) and the CBOE Volatility (VIX) Index as a proxy for 8 market sentiment. Three econometric models are considered to forecast SPY prices: (i) Auto 9 Regressive Integrated Moving Average (ARIMA), (ii) Generalized Auto Regressive Conditional 10 Heteroskedasticity (GARCH), and (iii) Vector Autoregression (VAR). These models are integrated 11 into two technical indicators, Bollinger Bands and Moving Average Convergence Divergence 12 (MACD), focusing on forecast performance. The profitability of various algorithmic trading 13 strategies are compared based on a combination of these two indicators.