Is Bitcoin on a Random Walking? – Alexandre Oliveira
In early 2019, Divya Aggarwal published the article “Do bitcoins follow a random walk model?”. The article investigates if BTC prices follow a random walk or not.
While many investors believe that there are opportunities in BTC trading, a rigorous analysis of the market efficiency has important meaning for the investment strategies suitable this market.
This post will summarize the article’s findings and analyze it’s implications for Bitcoin investors.
Is BTC Random Walking?
In statistical terms, this means:
H₀: Daily bitcoin returns follow a random walk
H₁: Daily bitcoin returns does not follow a random walk
To accomplish this, many statistical tests were performed including Durbin Watson (DW) statistics, the correlogram Q statistic and unit root tests.
The two chart below, showing the daily returns and squared returns of Bitcoin, strongly suggest that volatility is clustered.
Unit Root Tests (Stationarity)
Three tests including ADF, PP and KPSS LM statistic were done to check for presence of unit root without break. All the three tests confirmed that the series are stationary i.e. do not contain unit roots. Similarly, the innovational and additive outlier variants of the breakpoint unit root tests (intercept only and trend and intercept) were performed to examine the null hypothesis of the series containing a unit root. Again, all tests indicates that the BTC returns series is stationary.
ARCH
ARCH tests on pure random walk model for daily bitcoin returns are performed. The null hypothesis of absence of ARCH effect was rejected.
To examine the nature of volatility persistence in the random walk model of daily bitcoin returns, all ARCH-M, GARCH, TARCH and EGARCH models have been applied.
Results show that the volatility tends to persist for a longer time period along with having asymmetric effects. It also indicates that positive news have a larger impact on bitcoin returns in comparison to negative news.
- The multiple unit root tests and volatility persistence measures confirm that daily bitcoin returns do not follow a RWM
- The market inefficiency of bitcoin market is quite strong
- Significant positive asymmetric volatility reveal higher impact of positive news over negative news leading to persistent euphoria around positive news.
- Trading strategies may be created to explore the strong presence of volatility clustering in BTC returns.
Published at Sat, 08 Feb 2020 15:13:49 +0000
{flickr|100|campaign}
