On the determinants of bitcoin returns: a LASSO approach


 By Orestis Vravosinos (Economics ’18)

Orestis shares his recent work on bitcoin, joint with professors Theodore Panagiotidis (Univeristy of Macedonia) and Thanasis Stengos (University of Guelph).

The paper is in press and currently available online at Finance Research Letters. It first appeared as a working paper of the Rimini Centre for Economic Analysis (RCEA).

Extended abstract: Bitcoin has recently received increased attention by both investors and researchers. The consumer base, transaction frequency in the digital currencies market and the number of businesses and organizations accepting payments in bitcoin have been rapidly expanding.

In this paper, within a Least Absolute Shrinkage and Selection Operator (LASSO) framework, we examine the significance of factors such as stock market indices, exchange rates, gold and oil, central bank rates, internet trends and policy uncertainty as drivers of bitcoin returns for alternate time (sub)periods within 2010-2017. The advantage of LASSO is that it performs coefficient shrinkage (even setting some coefficients to zero) and in this way automates model selection. Search intensity and gold returns emerge as the most important factors for bitcoin returns.

References

Panagiotidis, T., Stengos, T. and Vravosinos, O. (2018). On the determinants of bitcoin returns: a LASSO approach. Finance Research Letters, doi: 10.1016/j.frl.2018.03.016