Predictability of stocks

Abstract

An important question in the financial market is the predictability of stock returns, which has essential and broad economic implications. Stock predictability is highly linked with the efficiency of the capital markets in allocating the available resources combined with their highly valued uses. This thesis analyzes the predictability of stock market returns in the USA. The aim of this research is to identify the factors that affect stocks and make them predictable. As a result, there is a pattern in the stock prices that traders can exploit and make profit from, i.e., beat the market. To find these factors, bivariate models are developed. The period of the analysis started from the 1st of January 1900 and ended up to 31st of December 2021. Regarding the results, there were some factors that could be used in order to predict stock prices. In addition, a homoscedasticity test is shown in order to examine the stability of the variance of the residual. Furthermore, an ARCH test is used to examine whether there is dynamic homoscedasticity. According to the results, the null hypothesis was rejected, that is, there was a strong dynamic heteroskedasticity. This research provides clear evidence of stock market return predictability in the United States of America using financial variables. Future research can be carried out to compare alternative ways to examine the stock predictability, such as splitting the period into sub periods for a better understanding of the relationship as well as including COVID-19 pandemic variables in order to see the effect of the pandemic on stocks.