Econometrica: Nov 2018, Volume 86, Issue 6

Long memory via networking

DOI: 10.3982/ECTA11930
p. 2221-2248

Susanne M. Schennach

Many time series exhibit “long memory”: Their autocorrelation function decays slowly with lag. This behavior has traditionally been modeled via unit roots or fractional Brownian motion and explained via aggregation of heterogeneous processes, nonlinearity, learning dynamics, regime switching, or structural breaks. This paper identifies a different and complementary mechanism for long‐memory generation by showing that it can naturally arise when a large number of simple linear homogeneous economic subsystems with short memory are interconnected to form a network such that the outputs of the subsystems are fed into the inputs of others. This networking picture yields a type of aggregation that is not merely additive, resulting in a collective behavior that is richer than that of individual subsystems. Interestingly, the long‐memory behavior is found to be almost entirely determined by the geometry of the network, while being relatively insensitive to the specific behavior of individual agents.



Log In To View Full Content

Supplemental Material

Supplement to "Long memory via networking"

This Supplement Material includes various extension of the paper’s main results, namely (i) deviations from power laws in the cn coefficients (ii) the presence of multiple sources of noise in the network (iii) the possibility of non integrable limiting power spectra and (iv) heterogeneity in the agents’ responses. It also includes the description of a simple and stylized variant of the Loss-Plosser model as well as a “toy” application based on the “input-output accounts” database compiled by the Bureau of Economic Analysis.

Read More View PDF


Supplement to "Long memory via networking"

This zip file contains the replication files for the manuscript.

Read More View ZIP


Back