Fractal connectivity networks of select stock returns exhibiting long range dependence

Abstract

Despite several attempts in applied econometrics and time series literature to identify the common channels contributing to fractal structures and long memory in multivariate financial time series, we propose a wavelet-based fractal connectivity analysis, which is the first application in economics and financial studies, enabling one to successfully segregate markets into fractally similar or diverse groups and find that developed markets have similar fractal structures. Similarly, stock returns of emerging markets exhibiting long-memory tend to follow similar fractal structures. Furthermore, network analyses of fractal connectivity support our findings on market efficiency which has theoretical roots in both fractal and adaptive market hypothesis.