Monetary regimes and statistical regularity: the Classical Gold Standard (1880-1913) through the lenses of Markov models

We aim at characterizing the Classical Gold Standard period (CGS) in order to verify if it is endowed with statistical regularity. We study the statistical properties of two-state annual transition matrices of countries switching from a sound state to a crisis state focusing on Reinhart and Rogoff 2009 dataset on external debt crises. The CGS period is governed by homogeneity both in time and across statistical units: the Homogeneous Markov Chain Model holds whereas the Mover Stayer Model does not. Our work is linked to the literature on the CGS and credibility (Bordo and Rockoff 1996). We follow a pure statistical approach to highlight two decisive channels of the credibility mechanism. The first is the stabilization of the probability of default of sound countries. The second is the fact that the CGS makes periphery/deficit countries homogeneous to the core with respect to the probability of default. Both channels are decisive because poor developing countries can borrow at favorable conditions and finance a level of investment greater than their capacity of saving.