Average diffusion of mean–variance from a network perspective: Evidence from the Chinese stock market

This paper investigates the movement of stocks in the \sigma-\mu plane from a mean perspective by constructing financial networks. The variation in diffusion entropy reveals that the network system tends to exhibit increasing stochasticity over time, which may provide a theoretical explanation for the failure of the mean–variance optimal portfolio. The results show that stocks with lower volatility and lower average returns may perform better in the future. Moreover, we propose a hedging strategy and enhance its performance by assigning portfolio weights based on the shortest path lengths from nodes to the barycenter.