A Model and Structural Analysis of Networked Bitcoin Transaction Flows

Min-Hsueh Chiu, Mayank Kejriwal

Abstract: Despite its unconventional origins, Bitcoin has emerged as the predominant cryptocurrency in the modern era and has entered mainstream discourse. Transactions in the Bitcoin ecosystem are different from those of ordinary finance, because of the way in which the cryptocurrency was designed. A proper structural understanding, and model, of Bitcoin transaction flows has largely been lacking. In this paper, we propose a model (based on directed acyclic graphs) that enables us to conduct structural analysis of networked Bitcoin transaction flows. Our model includes ‘activity’ measures, analogous to liquidity measures in ordinary financial markets, that are inspired by intuitions from thermodynamics. We apply the model and the activity measures to conduct structural analysis on a large transaction dataset from the early days of Bitcoin (first five years) when it was most in flux, and its future was still uncertain. Among other findings, our structural analysis suggests that the activity measure is correlated with major news events that affected Bitcoin. Our model could potentially be used to study other cryptocurrencies for which transaction data is available, as well as more recent Bitcoin transaction data.