Comparing GARCH and TGARCH Models for Predicting Prices Volatility of Three Popular Types of Cryptocurrencies

Shanhong Wang

《Frontiers in Economics and Management》 4卷1期 (2023/01) Pp. 1-27

https://doi.org/10.6981/FEM.202301_4(1).0001

摘要

This dissertation is the first to predict the price volatility of Bitcoin, Ethereum and Litecoin by employing both one‐day‐ahead and month‐ahead dynamic forecasting approaches. The predictive abilities of the ratio between the RMSFE (root mean square forecast error) of GARCH (1,1) (generalized autoregressive conditional heteroscedasticity) model and the RMSFE of TGARCH (1,1) (threshold generalized autoregressive conditional heteroscedasticity) model are evaluated and compared. Completeness checking of the comparative validity of the model is accomplished by the DM (Diebold‐Mariano) test. The parameters of the GARCH class models have first assessed by the in‐sample cryptocurrencies' closing price from August 7, 2015, to December 31, 2018. Out‐of‐sample price volatility prediction is operated recursively afterwards. The realized volatility of the three digital currencies in 2019 is utilized to compare the dynamic forecasting price volatilities depending on loss statistics (RMSFE). Then the RMSFE values are displayed following the ratios. The results are more biased in favour of a definition of cryptocurrencies as financial assets, which performs a certain difference from fiat money. Based on the performance of density function selection, the comparisons of RMSFE values and ratios, the TGARCH (1,1) model has better behaviour in out‐of‐sample price volatility forecasts with higher accuracy. Due to the unique nature of cryptocurrencies' price fluctuations, the results of future works with the same steps of this paper may change.

關鍵字

Bitcoin ； Ethereum ； Litecoin ； Modelling Variance ； Predictive Performance ； GARCH (1,1) ； TGARCH (1,1) ； Out‐of‐sample Forecasting ； Dynamic Predicting Method ； Optimization ； RMSFE ； DM Test

參考文獻

Al Guindy, M., 2021. Cryptocurrency price volatility and investor attention. International Review of Economics & Finance, 76, pp.556-570.

Almansour, B.Y., Alshater, M.M. and Almansour, A.Y., 2021. Performance of ARCH and GARCH models in forecasting cryptocurrency market volatility. Industrial Engineering & Management Systems, 20(2), pp.130-139.

Amiri, O., & Remita, M. R., 2022. Measuring the impact of covid-19 pandemic on the oil prices volatility by GARCH modeling. Advances in Mathematics: Scientific Journal, 11(6), pp.515–537.

Anwar, M., 2003. Islamicity of banking and modes of Islamic banking. Arab LQ, 18, p.62.

Barjašić, I., & Antulov-Fantulin, N. (2021). Time-varying volatility in Bitcoin market and information flow at minute-level frequency. Frontiers in Physics, 273.