Crypto network

The empirical literature has studied linkages in the cryptocurrency market because knowing how shocks pass from one currency to another helps policymakers and practitioners better counter their propagation in these and related markets. This paper contributes to this literature by proposing a methodology based on Granger causality and network analysis. Using the daily log- returns of 22 cryptocurrencies over the period 2018-2023, I develop a VAR model to infer unidirectional or bidirectional Granger causality among cryptocurrencies. These relationships are then transformed into a directed network and several centrality measures are calculated. The centrality measures are also observed over the years to understand the dynamics of the cryptocurrency network. I find out that each one unit increase in eigencentrality is associated with a 0.22 percent increase in log-returns. Cryptocurrencies are nontrivially connected, and in this sample Cardano, Dogecoin, Gridcoin, and Neo are amongst the most central in the network throughout the period. Some cryptocurrencies, such as Dogecoin or Neo, show decreasing centrality over the years, while others, such as Gridcoin, Litecoin, Namecoin, or Ripple, gain centrality. These results support the idea that the cryptocurrency market is no longer exclusively associated with Bitcoin and lay the groundwork for further study of shock propagation in financial markets.