This paper shows that shocks, in the presence of asymmetric propagation structures, diminish investors’ diversification benefits. First, we construct an interdependency network with assets as nodes, and links corresponding to cross-dependency in returns. Second, we show that higher heterogeneity in the structure of the network increases portfolio risk. In particular, diversification among assets with star-like network structures, where a central asset cross-affects other assets in the portfolio, results in the lowest level of diversification benefits. Finally, we empirically demonstrate that two distinct datasets of U.S. industries and international stock markets greatly resemble star-like network structures.

Practitioners allocate substantial resources to technical analysis whereas academic theories of market efficiency rule out technical trading profitability. We study this long-standing puzzle by applying a diverse set of machine learning algorithms. The results show that an investor can find profitable technical trading rules using past prices, and that this out-of-sample profitability decreases through time, showing that markets have become more efficient over time. In addition, we find that the evolutionary genetic algorithm’s attitude in not shying away from erroneous predictions gives it an edge in building profitable strategies compared to the strict loss-minimization-focused machine learning algorithms.