Fluctuations in interbank network dynamics

This work investigates the scaling properties of fluctuations in the flux of individual agents with respect to their average flux in an interbank network. The analyzed data provide information on daily values of f(i)(asset), the credit provided by bank i in the interbank network, and f(i)(liab), the credit received by bank i from the other banks of the network. The investigation focuses on the scaling properties of the fluctuations in the raw data f(i)(asset), f(i)(liab), and f(R,i)(ext)(t)=f(i)(asset)-f(i)(liab), as well as on similar properties internal and external fluctuations f(i)(int) and f(i)(ext), which are derived according to a recently proposed methodology [M. Argollo de Menezes and A. L. Barabasi, Phys. Rev. Lett. 93, 068701 (2004)]. Finally, a "rolling sampling" approach is introduced in order to deal with the nonstationarity of the fluxes. The results suggest that exponents are time varying, hinting that the considered interbank network is changing with time.