Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Volatility forecasting for cryptocurrencies under a heavy-tailed distribution

Vargas Pico, Diego Mauricio LU and Bylkova, Alina LU (2019) NEKN02 20191
Department of Economics
Abstract (Swedish)
In the recent years, cryptocurrencies have gained popularity and have experienced high price volatility. This essay pretends to examine how the multivariate GARCH models predict the volatility of these digital currencies and what implications exist if we consider the correlations among them to forecast their volatility. Using a bivariate Diagonal VECH and bivariate Diagonal BEKK this thesis checks the covariance on the returns of the two largest cryptocurrencies in terms of market capitalization, Bitcoin and Ethereum. Since the price returns of financial assets tend to have a non-normal behavior, we consider a distribution that allows the data to have heavier tails, assuming that this could be more realistic as it is observed in other... (More)
In the recent years, cryptocurrencies have gained popularity and have experienced high price volatility. This essay pretends to examine how the multivariate GARCH models predict the volatility of these digital currencies and what implications exist if we consider the correlations among them to forecast their volatility. Using a bivariate Diagonal VECH and bivariate Diagonal BEKK this thesis checks the covariance on the returns of the two largest cryptocurrencies in terms of market capitalization, Bitcoin and Ethereum. Since the price returns of financial assets tend to have a non-normal behavior, we consider a distribution that allows the data to have heavier tails, assuming that this could be more realistic as it is observed in other financial markets. After employing the different models, we find that the best-fit distribution for estimating the conditional (co)variance is the Student’s t, which allows the data to have fatter tails. We find that GARCH(1,1) is the best specification for the conditional (co)variance of Bitcoin and Ethereum and when performing the forecast we find that the best model is estimated with the Diagonal VECH approach. (Less)
Please use this url to cite or link to this publication:
author
Vargas Pico, Diego Mauricio LU and Bylkova, Alina LU
supervisor
organization
course
NEKN02 20191
year
type
H1 - Master's Degree (One Year)
subject
keywords
Cryptocurrency, Bivariate Diagonal BEKK, Bivariate Diagonal VECH, MGARCH, Volatility
language
English
id
8983277
date added to LUP
2019-08-08 10:28:49
date last changed
2019-08-08 10:28:49
@misc{8983277,
  abstract     = {{In the recent years, cryptocurrencies have gained popularity and have experienced high price volatility. This essay pretends to examine how the multivariate GARCH models predict the volatility of these digital currencies and what implications exist if we consider the correlations among them to forecast their volatility. Using a bivariate Diagonal VECH and bivariate Diagonal BEKK this thesis checks the covariance on the returns of the two largest cryptocurrencies in terms of market capitalization, Bitcoin and Ethereum. Since the price returns of financial assets tend to have a non-normal behavior, we consider a distribution that allows the data to have heavier tails, assuming that this could be more realistic as it is observed in other financial markets. After employing the different models, we find that the best-fit distribution for estimating the conditional (co)variance is the Student’s t, which allows the data to have fatter tails. We find that GARCH(1,1) is the best specification for the conditional (co)variance of Bitcoin and Ethereum and when performing the forecast we find that the best model is estimated with the Diagonal VECH approach.}},
  author       = {{Vargas Pico, Diego Mauricio and Bylkova, Alina}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Volatility forecasting for cryptocurrencies under a heavy-tailed distribution}},
  year         = {{2019}},
}