Potential drivers of Bitcoin returns: evidence within a LASSO framework

Abstract

This thesis examines the relationship between bitcoin returns and 36 potential explanatory variables. We first present in brief how bitcoin operates, its history, and its uses. Afterward, we present an extensive review of the literature. We use a GARCH filter in order to deal with the effects of volatility clustering in our models. To capture the effects of technical factors, market forces of bitcoin’s supply and demand, investors’ attractiveness to bitcoin, and global macroeconomic and financial indicators, we employ a LASSO regression due to its ability to perform both variable selection and regularization in the analysis. We estimate models for the period 2013-2020 (full sample), and 3 sub-periods of it separately. Our results reveal policy uncertainty and exchange rates as the most consistent, important drivers of bitcoin returns.