PurposeThis study aims to contribute to the current and well-documented phenomenon of hysteresis in the US economy after the Great Financial Recession. Compelling reasons lead me to believe that Verdoorn's law can be used to explain this phenomenon by relating hysteresis to a fall in the size of returns to scale of the whole economy.Design/methodology/approachVerdoorn's law is estimated using a time-varying regression model that employs Bayesian methods to examine the evolution of the Verdoorn coefficient. The investigation uses a bivariate time-varying model to estimate long-run growth rates of output and productivity while controlling for potential endogeneity problems.FindingsThe study finds substantial variation in the Verdoorn coefficient across time as well as a significant fall of it in the onset of the Great Financial Recession, which confirms the presence of hysteresis in the US economy. Additionally, it also finds a fall in the size of productivity shocks, and in the long-run growth rates of output and productivity according to previous studies.Originality/valueThe empirical investigation uses, novel to the literature on Verdoorn's law to date, a bivariate time-varying regression model with stochastic volatility. It also employs quarterly productivity data for a considerably long period, which to my knowledge, has not been used by previous works. Most importantly, this approach contemplates positive hysteresis––typically neglected––which provides a less bleak panorama.

This paper investigates whether a unit root process and nonlinearities can characterize real commodity prices for six major primary goods. An unconstrained two-regime threshold autoregressive model is used with an autoregressive unit root. Among the main results, it is found that terms of trade for agricultural, mineral, non-tropical, and non-oil goods are nonlinear processes that are characterized by a unit root process. The finding of nonlinearities explains why the deterioration of the terms of trade has been episodic. Additionally, we found there is no statistical evidence supporting the Prebisch-Singer hypothesis for agricultural, mineral, non-tropical, and non-oil goods.