Determinants of Bitcoin price ๐
-
The introduction is well-written, but there are a couple of questions which I’d like to see answered:
- You mention that Bitcoin is a “speculative asset”, but why exactly can it not be explained by traditional asset pricing models?
- Are there any other predictors other than social media sentiment?
- Where is the logic underlying the link with the pandemic? Why should the “fundamental” relationship change in the pandemic?
- Think about a conceptual framework here
-
I like your contrasting of Bitcoin with gold, that makes it more suitable to a general audience of finance readers (as opposed to crypto-enthousiasts!)
-
The discussion in section 2.2 is very relevant and key to your empirical design - I like the aspects you mention, but you should pay more attention to the “standard” demand-supply story. Supply is basically exogenous, and (almost perfectly) anticipated, which you can show, and therefore, it is useful to focus on demand.
-
Then, you should talk about various demand factors, among which are some you will research empirically.
-
Paragraph 2.3 and 2.4 do not really seem to have a clear function
-
2.6 seems to explain that sentiment can be something which (imo) determines demand
-
In your hypotheses (p. 16), what do you mean by relationship? Correlation?
-
You should recap your hypotheses in the methodology section for readability
-
The econometric approach has to be thought through: what is the frequency of returns that you measure? Daily?
- Then, you have to think about the temporary dimension, maybe estimate an ARMA\VAR model
- Rigourous alternative: Instrument sentiment by something else, but might be difficult
- Therefore: confine to non-causal analysis
- There is a couple of $\beta$’s in the specification that are superfluous
-
There should also be a general discussion about endogeneity (sentiment creates returns, but returns also create sentiment)
-
Potentially complicated time-series relationship