Estimating the New Keynesian Phillips Curve (NKPC) with Fat-tailed Events

This paper provides estimation of the New Keynesian Phillips curve accounting for the unexpected large shocks such as Covid-19. The recent pandemic distorted the estimates of the output gap derived using the regular trend cycle decomposition of GDP (HP Filter, BP Filter, Kalman Filter). We propose a modified unobserved components model (UCM) by introducing an additional Student-t distributed irregular component in the trend cycle decomposition of GDP, which successfully isolates transitory shocks like COVID-19 from trend and cycle estimates. We also construct a model-based measure of inflation expectations that captures adaptive learning from a long inflation history and real-time updating during the pandemic. For India, we find a stable linear NKPC. Our results demonstrate that accounting for fat-tailed events is crucial for obtaining reliable Phillips curve estimates in emerging markets.