A statistical-mathematical analysis of the macroeconomic effects of long-memory total factor productivity

In this manuscript, the authors depart from the conventional assumption that total factor productivity is well fitted by a short-memory process and show that, in Japan and other selected countries, this variable admits a long-memory representation, implying that the technology shocks hitting their respective economies are more persistent compared to those predicted by an ARMA(p,d,q) process. Then, in order to assess the macroeconomic effects of such persistence, the authors write down a Real Business Cycle model in which total factor productivity is an ARFIMA(p,d,q) model and compute the related impulse response functions. The results obtained by the authors indicate that, in the context of the standard Real Business Cycle model, the long-memory property of total factor productivity is an important propagation mechanism for both technology and fiscal shocks that reproduces the higher-order moments of macroeconomic time series.