Bubbles everywhere: are cryptocurrencies and technological stocks well explained by a causal-noncausal bubble model?

Abstract

This master thesis studies the dynamics of cryptocurrencies and technological stocks in the recent years. The aim was to understand if these assets could be modelled using a recently developed model called the Mixed Causal-Noncausal model and if their bubbly behavior could be explained through the use of this model. Cryptocurrencies and technological stocks are not the most understood assets on the market and this works provides newer insights on their features and patterns. We also try to investigate if some common macroeconomic factors, market indices and other assets could have an impact on our assets. Our first contribution is to asses if particular assets such as cryptocurrencies or technological stocks behavior could be explained and confirmed by the model we use. Our second contribution is to provide forecasts based on the models we obtain and to assess the performance of such forecasts. This analysis results in new insights on the cryptocurrencies and technological stocks. Firstly, we identify a strong relationship between all the cryptocurrencies under the scope of this thesis and gold returns. We also identify a significant relationship between S&P 500 and our cryptocurrencies suggesting that they could behave either as safe-haven when the traditional markets are volatile and as speculative instruments when they are calm. The results obtained for the technological stocks are less homogenous but we can assume that most of them do have explosive roots and behave as bubbles according to our modelling procedure. We also identify significant relationships between Gross Domestic Product and Crude Oil and US Treasury Bond returns for most of these assets. Finally, the forecasting performances of our models is somehow mitigated. We are able to identify trends for some of our assets and not for others. However, we can conclude that the mixed models had good performance in identifying the trend, especially in a short-term horizon. We believe that Mixed Causal-Noncausal models could be used in order to put in place financial strategies when encountering bubbles. Putting in place momentum strategies or use it to hedge a portfolio is something that can be considered based on the results obtained.