Bitcoin Community Models

---

The final chapter of my thesis, it compares the Bitcoin transaction network with that of simulated networks from specific random graph models. In particular, taking the daily community level graph. We compare the observed data to a menu of random graph models, including Erdös-Renyi and preferential attachment amongst others.

The comparison is done by testing differences in network structure using Auerbach (2022). This contributes to the literature by providing an empirical piece on Bitcoin network formation using the latest econometric methods.

This approach identifies which model, out of the menu, is the closest to the realised transaction network for each day. In doing so we can say which models are definitely not good specifications and which mechanisms are most likely to be taking place. By repeating this exercise daily, we observe evolution in network formation, and relate it to price behaviour.