Table 3 Summary of prior literature on Bitcoin reaction to investors’ sentiment and various news
Author(s) | Period | Variables | Quantitative methods | Empirical outcomes |
|---|---|---|---|---|
Aharon et al. (2020) | January 1, 2011–July 4, 2020 | Bitcoin, Ethereum, Bitcoin-cash, Ripple, Twitter Market Uncertainty Index, Twitter Economic Uncertainty Index | OLS, GARCH, Granger-causality in distributions | Strong causal connection among the social media uncertainty and cryptocurrency returns |
AlNemer et al. (2021) | January 15, 2013–November 15, 2020 | Bitcoin, Dogecoin, Ethereum, Litecoin, Tether, Sentix Investor Confidence | Wavelet coherency analysis | Long-term positive connection between Bitcoin prices and Sentix Investor Confidence |
Aslanidis et al. (2022) | August 7, 2015–April 22, 2021 | Bitcoin, Google Trends Cryptocurrency Attention Index (GTC) | Transfer entropy | Two-way stream of information among GTC and cryptocurrency returns up to six days |
Banerjee et al. (2022) | January 1, 2020–April 15 2021 | Top 30 cryptocurrencies by market capitalization, RavenPack COVID-19 sentiments | Transfer entropy | The connection between COVID-19 news sentiment and cryptocurrency returns is nonlinear |
Bonaparte and Bernile (2022) | January 2004 –March 2022 | Bitcoin, Ethereum, BNB, Cardano, Solano, Terra, Dogecoin, Crypto Regulation Sentiment Index (CRSX) | Regression analysis | There is no statistically significant long-term effect of CRSX on the price of cryptocurrencies |
Bourghelle et al. (2022) | January 21, 2020–May 25, 2021 | Bitcoin, fear and greed index | Linear and nonlinear vector autoregressive (VAR) model | The impact of market sentiment depends on time |
Bouteska et al. (2022) | January 1, 2015–September 30, 2022 | Bitcoin, Cryptocurrency Index (CRIX), Volatility CryptoIndeX (VCRIX), sentiment measures based on StockTwits and Reddit | Principal component analysis (PCA) method, vector autoregressive model (VAR) | Investor sentiment, as assessed by messages pertaining to the financial aspect of cryptocurrencies, has a greater predictive ability and yields better outcomes than the cryptocurrency index, particularly during times of market turmoil |
Burggraf et al. (2021) | April 2013–February 2019 | Logarithmic Bitcoin returns, microeconomic and macroeconomic financial and economic attitudes revealed by search (FEARS) | Transfer entropy, threshold regression, ordinary least squares (OLS), generalized least squares (GLS), two-stage least squares (2SLS) regressions, VAR-Granger analysis | The effect of investor emotion on Bitcoin return is adverse and statistically significant |
Chen et al. (2020) | January 15, 2020–April 24, 2020 | Bitcoin price dynamics, VIX, Google Trends | Vector autoregressive (VAR) models | Growing fear of the coronavirus leads to negative Bitcoin returns and high trading volume |
Ciaian et al. (2016) | November 2009 –May 2015 | Bitcoin price, the volume of daily Bitcoin views on Wikipedia, new members and new posts on online Bitcoin forums | Vector error correction (VECM) model | The arrival of fresh news positively influence Bitcoin price |
Dias et al. (2022) | January 1, 2017–December 31, 2021, | Bitcoin, Google search volume, Twitter happiness index, Wikipedia page views, news sentiment, VIX, daily merits shared in bitcointalk.org | Principal component analysis, Quantile regression approach | Bitcoin returns can be significantly predicted by investor interest and sentiment |
Fang et al. (2020) | May 2013–May 2019 | Bitcoin, Ethereum, Ripple, Litecoin, New Economy Movement, Global Economic Policy Uncertainty (GEPU), of News-based Implied Volatility (NVIX) | GARCH-MIDAS | Negative impact of NVIX on the cryptocurrencies’ long-term volatility |
Figà-Talamanca and Patacca (2020) | January 2012–December 2018 | Bitcoin returns and trading volume, Google Search Volume Index (SVI) | Vector AutoRegressive (VAR) model | Bitcoin returns are not affected by trading volume and SVI |
French (2021) | October 1, 2013–September 15, 2020 | Bitcoin, Twitter-based market uncertainty index | Pairwise Granger causality, Bayesian vector auto-regression (BVAR) | A significant predictor of Bitcoin returns only throughout the COVID-19 period is the Twitter-based market uncertainty index |
Gaies et al. (2021) | August 2011–July 2020 | Bitcoin Misery Index, VIX, the Kansas City Financial Stress Index, the 10-year US nominal interest rate | Nonlinear autoregressive distributed lag model | An increase in the level of optimistic (pessimistic) sentiment has a positive (adverse) influence on Bitcoin returns |
Gök et al. (2022) | June 1, 2011–August 30, 2021 | Bitcoin, gold, US10 year Treasury notes, Twitter-based economic uncertainty index, geopolitical risk index, US VIX, daily infectious disease equity market volatility tracker | Causality-in-quantiles, Wavelet decomposition | Causality-in-variance from Twitter-based economic uncertainty to Bitcoin |
Guégan and Renault (2021) | August 2017–December 2019 | Bitcoin prices and retuns, StockTwits sentiment | Multivariate regressions, Granger causality tests | Investor sentiment predict Bitcoin returns for high frequencies (up to 15 min) |
Güler (2021) | February 2018–August 2020 | Bitcoin trading volume, Crypto Fear & Greed Index, Weekly American Association of Individual Investors Index (AAII) | GARCH models, Vector autoregressive (VAR) model | Both rational and irrational investor sentiments influence Bitcoin returns |
Havidz et al. (2022b) | March 18, 2021–August 31, 2021 | Bitcoin, Ethereum, Vaccine confidence index, Global fear index, Panic index, Sentiment index, blockchain features | Autoregressive distributed lag (ARDL) | Negative connection among Global fear index and Bitcoin returns Vaccine confidence index and Global fear index were insignificant to Ethereum returns Panic index and Sentiment index were insignificant in the long run to Bitcoin and Ethereum returns |
Jin et al. (2021) | July 9, 2012–June 24, 2013 | Daily Bitcoin price | Empirical mode decomposition with adaptive noise (CEEMDAN)-based event analysis | The announcement of 2013 Cyprus bailout substantially intensified the strength of short-term oscillations in Bitcoin prices |
Kim et al. (2022) | November 2017–April 2018 December 2018–May 2019 | Trading volume and closing price of Bitcoin, number of Bitcoin searches on Google, number of positive/negative sentiments about Bitcoin | Hidden Markov model (HMM) | Optimistic social sentiment is more relevant throughout a bullish trend, whilst adverse social emotion is more significant during a bearish market |
Kyriazis et al. (2022) | January 1, 2020–July 25, 2021 | Bitcoin, Ethereum, Binance Coin, Cardano, Ripple, Dogecoin, Bitcoin Cash, Litecoin, Ethereum Classic, Stellar, economic and market uncertainty indices | Linear and nonlinear Granger causality tests | Twitter sentiment is noticed to have a significant impact on investigated cryptocurrencies |
Li et al. (2022) | January 2012–October 2021 | Bitcoin returns, Google News | Vector autoregressive (VAR) framework | In the bubble period, media coverage (whether positive or negative) has a positive link with Bitcoin returns the following day, but there is no significant relation in the post-bubble period |
Li et al. (2021b) | January 2013–April 2019 | Bitcoin return, domestic and foreign events | GARCH-X model | Domestic events positively influence Bitcoin price volatility, whereas foreign events impact both BTC price return and volatility |
Lyócsa et al. (2020) | January 2013–December 2018 | Bitcoin, news about the regulation of Bitcoin, hacking attacks on Bitcoin exchanges, investor sentiment, macroeconomic news | Quantile regressions | Increased Bitcoin volatility a day ahead of publication an article towards Bitcoin regulation |
Mahdi and Al-Abdulla (2022) | January 3, 2020 −September 1, 2021 | Bitcoin, gold, RavenPack coronavirus news-based indices | Quantile-on-quantile regression model | The distribution of Bitcoin returns is affected asymmetrically by positive and negative shocks in coronavirus-related news |
Mai et al. (2018) | January 1, 2012–December 31, 2014 | Bitcoin price, returns, trading and transaction volume, number of positive and negative posts, number of positive and negative tweets | Vector error correction (VECM) model | Social media sentiment can explain and predict Bitcoin value |
Mokni et al. (2022) | January 2, 2018–December 10, 2020 | Bitcoin, the fear and greed index | Symmetric and asymmetric causality analysis, quantile autoregressive regression model | The Bitcoin price has a strong impact on investor sentiment |
Naeem et al. (2021b) | March 7, 2016–December 29, 2019 | Bitcoin, Litecoin, Ripple, Dash, Monero, Ethereum, Twitter Happiness index, FEARS index | OLS, quantile regression (QR), cross-quantilogram (CQ) | Cryptocurrency returns are determined more by sentiment spread over social media than with macroeconomic news |
Nair (2021) | September 1, 2018–April 30, 2021 | Bitcoin, Ethereum, Litecoin, Neocoin | Decomposition of returns, GARCH framework, vector autoregressive model | The response of crypto markets to negative news is equivalent to how they respond to good news |
Philippas et al. (2019) | January 1, 2016–May 28, 2018 | Bitcoin prices, Twitter and Google Trends | Dual process diffusion model | Media networks have only a limited impact on Bitcoin prices, which is larger on periods with greater incertitude |
Polat et al. (2022) | January 1, 2019 –January 31, 2021 | Bitcoin, Thomson Reuters MarketPsych Indices | Bivariate vector autoregressive (VAR) models | A rise in fear sentiment has a longer and more significant adverse effect on Bitcoin returns |
Rajput et al. (2020) | January 2013–December 2018 | Bitcoin sentient index, Bitcoin returns, volume traded and volatility | Linear and nonlinear autoregressive distributed lag (ARDL) models | Positive connection of Bitcoin sentiment index with its returns and volume, but a negative connection with its return volatility |
Raza et al. (2022b) | January 2016–March 2021 | Bitcoin, Dash, Ethereum, Litecoin, NEM, Ripple, Google Trends data | Causality-in- quantiles test | The price of cryptocurrencies can be accurately predicted by using Google Trends |
Rognone et al. (2020) | January 1, 2012–November 1, 2018 | Bitcoin, Forex, Sentiment indices | Exogenous vector autoregressive (VAR-X) model | Bitcoin reacts positively to both positive and negative news, but cyber-attack and fraud news lessen its returns and volatility |
Sabah (2020) | February 9, 2014–December 31, 2018 | Venues that accept cryptocurrencies as a payment method, market capitalization and market cap Weighted Cryptoz Index Volatility for top 10, 25, 50 and 100 cryptocurrencies | Regression analysis, bivariate vector autoregression, Granger causality | Investor attention as measured through the number of new business venues that accept cryptocurrencies as a form of payment is a predictor of crypto volatility |
Salisu and Ogbonna (2021) | September 2, 2019–September 29, 2020 | Gtrend, Bitcoin, Ethereum, Litecoin, Ripple | GARCH-MIDAS | Fear-generated news set off by the COVID-19 pandemic boosts the return volatilities of the cryptocurrencies contrasted with the period prior to the contagion |
Tong et al. (2022) | January 1, 2017–January 26, 2022 | 24 cryptocurrencies, Google Trends, daily numbers of Twitter tweets | Transfer entropy | Twitter has a higher information flow toward cryptocurrencies than the other way around |
Vurur (2021) | January 8, 2020–December 31, 2020 | Bitcoin, Ethereum, Ripple, Panic index | Autoregressive-Distributed Lag Cointegration, Hatemi-J asymmetric causality | Rises in the Panic index diminish the cryptocurrencies’ value |
Wu et al. (2021b) | August 9, 2015–July 7, 2020 | Bitcoin, Ethereum, Litecoin, Ripple, Twitter-based economic uncertainty and Twitter-based market uncertainty | Granger causality test using the recursive evolving window approach | Variations in the Twitter-based economic policy uncertainty (EPU) indices are positively connected to the cryptocurrencies’ returns during the COVID-19 period |
Xia et al. (2022) | September 19, 2014–May 20, 2022 | Bitcoin, Economic Policy Uncertainty (EPU) and Cryptocurrency Uncertainty (UCRY) indices | GARCH-MIDAS | Global economic policy uncertainty has a significant adverse impact on the long-term volatility of Bitcoin, whereas cryptocurrency uncertainty has a beneficial effect |
Zhang et al. (2022) | January 1, 2020–September 18, 2020 | Crude oil, gold, Bitcoin, RavenPack specific COVID-19 news-related indices | Time–frequency analysis method | Panic sentiment and media hype influence Bitcoin |