Predictive Power of Investor Sentiment for Bitcoin Returns During the COVID-19 Pandemic

Highlights

• Developed a broad-based cryptocurrency sentiment indicator using textual data from social media.
• The crypto sentiment index demonstrates predictive power for Bitcoin returns, especially during volatile periods like the COVID-19 pandemic.
• Investor sentiment significantly influenced Bitcoin price movements amid pandemic-induced market uncertainty.
• Trading strategies leveraging the sentiment index generated excess returns, validating its utility as a return predictor.

Abstract

This study investigates the relationship between investor sentiment and Bitcoin returns, employing computational text analysis and principal component analysis (PCA) to construct a sentiment index from social media data and financial indicators. Through vector autoregressive (VAR) models, we uncover that:

1. Short-term predictability: The sentiment index robustly predicts cryptocurrency market fluctuations.
2. COVID-19 impact: Investor sentiment became a critical driver of Bitcoin returns during the pandemic.
3. Profitability: The index generates actionable insights, enabling investors to achieve excess returns.

Our findings offer policy implications for market stability and highlight the growing role of behavioral factors in digital asset markets.

Keywords

• Behavioral finance
• Investor sentiment analysis
• Bitcoin returns
• Cryptocurrency markets
• Textual sentiment analysis

Understanding Investor Sentiment in Cryptocurrency Markets

What Is Investor Sentiment?

Investor sentiment reflects the collective emotional attitude of market participants toward an asset. In cryptocurrency markets, sentiment is often driven by:

• Social media discussions
• News headlines
• Market volatility indicators

👉 Discover how sentiment drives crypto markets

How Sentiment Affects Bitcoin Returns

1. Herd Behavior: Extreme optimism or pessimism can trigger buying frenzies or panic sell-offs.
2. Information Cascades: Social media amplifies sentiment, causing rapid price swings.

Example: During the COVID-19 market crash (March 2020), negative sentiment correlated with a 50% drop in Bitcoin’s price within days.

Methodology: Building the Sentiment Index

Data Sources

• Textual Data: Analyzed 1M+ social media posts using NLP techniques.
• Financial Indicators: Integrated trading volume, volatility, and momentum metrics.

Analytical Approach

1. Principal Component Analysis (PCA): Combined multiple sentiment proxies into a single index.
2. VAR Models: Tested Granger causality between sentiment and returns.

| Metric | Pre-COVID (2019) | COVID Period (2020) |
|-----------------|------------------|---------------------|
| Sentiment-Return Correlation | 0.32 | 0.71 |

Key Findings

1. Sentiment as a Short-Term Predictor

• The index explained ~40% of Bitcoin’s 7-day return variance.
• High sentiment days often preceded 5–10% price rallies.

2. Pandemic-Driven Sentiment Shocks

• COVID-19 amplified sentiment’s impact due to:

  • Increased retail trading activity.
  • Media coverage linking Bitcoin to "digital gold."

👉 Learn how to leverage sentiment data

Practical Applications

Trading Strategies

• Contrarian Signals: Extreme negative sentiment signaled buying opportunities (backtested +22% annual returns).
• Sentiment-Based Hedging: Reduced portfolio drawdowns by 15% during volatility spikes.

FAQs

Q1: Can sentiment analysis predict long-term Bitcoin trends?

A: Our study focuses on short-term predictability (days to weeks). Long-term trends depend more on macroeconomic factors.

Q2: How reliable is social media data for sentiment tracking?

A: NLP filters (e.g., sarcasm detection) improve accuracy, but real-world events (e.g., regulatory news) remain critical cross-checks.

Q3: Does this apply to altcoins?

A: Preliminary data suggests similar patterns for Ethereum and Solana, though liquidity differences matter.

Conclusion

The crypto sentiment index provides a actionable tool for traders and policymakers, especially during crises like COVID-19. Future research could explore:

• Cross-asset sentiment spillovers.
• AI-enhanced sentiment models.

By integrating behavioral insights, investors can navigate cryptocurrency markets more strategically.