Comparative effectiveness of the ARMA-GARCH over simple strategies

Abstract

This End-of-Degree project is intended to assess if the combination of two different statistical models —ARMA and GARCH— can perform better than the common buy and hold strategy, as well as other simple strategies. There are two different reasons for me to write this paper. The first of them is an academic motivation to apply the studies carried out in a particular course I enrolled in, Forecasting Techniques, and test the accuracy of the models learned in the field of Finance. The second is a personal motivation to deep into the financial markets and assess the difficultness of investing in them. The project as such will firstly consist of an introduction where we will debate opposed ideas on the notion of the Efficient Market Hypothesis, then we will discuss the different types of existing analytical methods for understanding and investing in those markets and, finally, we will delve into the empirical part, where we will be using R to test the performance of all the models. In the end, we will have enough proofs to ascertain that the ARMA-GARCH outperforms the rest, yet there will be some caveats to it.