Andrajit Dube: A Note on Debt, Growth, and Causality:
Abstract: This note documents the timing in the relationship between the debt-to-GDP ratio and real GDP growth in advanced economies during the post World War II period using the Reinhart and Rogoff dataset. I first show that the debt ratio is more clearly associated with the 5-year past average growth rate, rather than the 5-year forward average growth rate–indicating a problem of reverse causality. Indeed, there is little evidence of a lower growth rate above the 90 percent threshold when using the 5-year forward average growth rate.
I use a number of simple tools to account for some of the reverse causality in the bivariate regression–such as using forward growth rates, instrumenting the current debt ratio with its lag, and controlling for lagged GDP growth rates. These simple methods of accounting for reverse causality diminish the size of the association by between 50 and 70 percent, with the linear regression estimate indistinguishable from zero. Finally non- and semi-parametric plots provide visual confirmation that the relationship between debt-to-GDP ratio and growth is essentially flat for debt ratios exceeding 30 percent when we (1) use forward growth rates, (2) control for past GDP growth, or both.