Hi,
I am trying to forecast bitcoin volatility with the use of Google trends data. The model I am using is a GARCH (1.1) My total time period is 1.1.2014-28.9.2018, while my estimation period is 1.1.2014-31.12.2016, and forecast period is 1.1.2017-28.9.2018.
Below you can find my results for estimation period:
Based on the results I perform a forecast :
What is the above graph showing? It seems that it has forecasted a constant price for the whole forecast period of 1.1.2017-28.9.2018 and is therefore just a straight line because return= 0%.
However the line below "forecast of variance" show the forecast of the variance. What I find strange is that the forecast of the variance is showing pretty much identical trendline to the realized Bitcoin price graph, ie real world Price development of Bitcoin. The estimation equation does not include data for 1.1.2017-28.9.2018 and only uses 1.1.2014-31.12.2016, therefore I find it strange that the forecast of variance is too similar to the realized Bitcoin prices 1.1.2017-28.9.2018.
Below is the estimation equations where i have put the dependent variable Bitcoin_ln in the mean equation followed by a constand. In the variance equation I have the variable for daily google trend search data. Am I somehow inputting this equation wrong? If so maybe that is the explanation of the forecast results. Lastly I have also showed a screencap how I have entered the forecast specifications. Is there anything I need to write in the GARCH(optional) part.
Bitcoin volatility forecast (question with pictures in attachment)
Moderators: EViews Gareth, EViews Moderator
-
- Posts: 9
- Joined: Wed Apr 22, 2015 7:57 am
-
- Posts: 9
- Joined: Wed Apr 22, 2015 7:57 am
Return to “Econometric Discussions”
Who is online
Users browsing this forum: No registered users and 13 guests