Are stock market networks non-fractal? Evidence from New York Stock Exchange

- We construct the NYSE network by using minimal spanning tree (MST).
- The (non)-fractal property of the stock market network is studied by using the edge-covering with simulated annealing method.
- The empirical results show that the degree of stocks obeys power-law distribution.
- Our finding suggest that the highly connected stocks connect with each other directly, i.e., the stock market network is non-fractal.

In this paper, we investigate the fractal (non-fractal) property of stock market network by using the edge-covering with simulated annealing method. We choose the daily closing price of 2109 stocks traded on the NYSE during the period from 2011 to 2014 as dataset and construct the network by using minimal spanning tree (MST). The empirical results show that the degree of stocks obeys power-law distribution and the highly connected stocks connect with each other directly, i.e., the stock market network is non-fractal. Our work provides a new perspective on risk management, which can be used in other network-based financial systems.