Credit, output and financial stress: A non-linear LVSTAR application to Brazil

Abstract

We estimate a bi-variate LVSTAR model to investigate the non-linear interplay between Brazil’s credit, output and financial cycles between 1999 and 2017. We use financial stress as the regime-switching variable to assess how it impacts economic performance. We find evidence of a lengthy transition between regimes. Moreover, Granger Causality tests indicate statistical precedence of credit growth through financial intermediaries. After credit shocks, we find non-mean-reverting trajectories and that financial stress worsens the economic downturn. Counterintuitively, total credit growth increases after adverse output shocks. This appears to be caused by elevated public credit provision relative to private credit flows, triggered by anticyclical policy.