A non-linear analysis of Gibson’s paradox

Abstract

In contrast to the previous literature, this paper adopts a multivariate, nonlinear framework to analyse Gibson’s paradox in the Netherlands over the period 1800-2012. Specifically, SSA (Singular Spectrum Analysis) and MSSA (Multichannel Singular Spectrum Analysis) techniques are used. It is shown that changes in monetary policy regimes or volatility in the price of gold by themselves cannot account for the behaviour of government bond yields and prices over the last 200 years. However, the inclusion of changes in the real rate of return on capital, M1, primary credit rate, expected inflation, and money purchasing power enables a nonlinear model to account for a sizeable percentage of the total variance of Dutch bond yields.