Empirical analysis of a debt-augmented Goodwin model for the United States

Abstract

The Goodwin–Keen model was introduced to study the structural instability of debt-financed economies. Despite its theoretical appeal, no empirical study has focused on this model to date. Using u.s. data for non-financial firms over the period 1959–2019, this paper tests the empirical validity of a Goodwin–Keen model accounting for dividends. We propose a two-step procedure to simultaneously estimate parameters and quantify their uncertainty. The model satisfactorily captures the historical cycles in the wage share and employment rate, while reflecting the trend growth in the debt-to-output ratio. The estimation provides meaningful parameters, although their large uncertainty suggests that the model fails to fully account for the private debt dynamics. The estimated model assigns a probability of at most 11% to the occurrence of a private-debt overhang over the 21st century. However, it would have failed to foretell the Global Financial Crisis. Further work is therefore needed to improve its prospective capacity.