PREDICITING BULL AND BEAR IN S&P500

Abstract: This paper seeks to investigate if factors extracted from macroeconomic and financial variables can improve the forecast accuracy of the bull and bear market in the S&P500 stock index. The study extended the models constructed by Nyberg(2012) and Chen (2009) by augmenting their model with factors. Very little, if any research has been done in modelling the bull and bear using this approach. After using the Bry-Boschan method to identify the two regimes in the stock market, eleven models were constructed using a static probit or dynamic probit model framework. The out of sample forecast results indicates that, probit models augmented with factors have a relatively lower Quadratic Probability Score (QPS) than the corresponding models without factors. Among all the models employed, the dynamic probit model outperformed all the models, while the static probit model without factors is the least performing model. The results also showed that returns on assets and money stock are among the key leading indicators of the S&P500 stock index. Thus, there is evidence that factors can improve the forecast accuracy of the bull and bear market in the S&P500 stock index.