Factors Influencing Cryptocurrency Prices: Evidence from Bitcoin, Ethereum, Dash, Litecoin, and Monero

Battle Count: 83

The paper provides valuable insights into cryptocurrency price determinants (market beta, trading volume, volatility, attractiveness) that could inform trading strategies. The finding that Bitcoin and Ethereum have higher market beta (0.79 and 0.38 long-run) and faster adjustment speeds (23.68% and 12.76%) is relevant for cross-asset hedging and pairs trading. The negative short-run SP500-Bitcoin relationship (-0.20) suggests potential mean-reversion or hedging opportunities. However, the paper is primarily descriptive/explanatory rather than predictive, and does not develop trading signals or backtest strategies. The ARDL framework is more suited for understanding relationships than generating real-time trading signals.

The ARDL bound testing approach is a well-established econometric technique available in standard software (Eviews, Stata, R via 'ardl' package). The main complexity lies in: (1) constructing the Crypto 50 index from 50 constituent cryptocurrencies with market-cap weighting; (2) proper lag selection using SIC; (3) determining appropriate ARDL case specification (I-IV) for each cryptocurrency; (4) handling mixed I(0)/I(1) integration orders; (5) applying HAC-robust standard errors. The methodology is moderately complex for an econometrician but accessible with standard software. No machine learning or deep learning components are involved.

The paper uses publicly available data sources (BitInfoCharts, Yahoo Finance, World Bank, Google Trends, CoinMarketCap) and standard econometric software (Eviews 9.0). The ARDL methodology is well-documented. However, the custom Crypto 50 index construction methodology, while described, would require significant effort to replicate exactly. No code or supplementary materials are provided. The weekly data frequency and specific sample periods for each cryptocurrency (varying N values) add complexity to replication.