Primo Beekhuis

  School of Business and Economics | Bachelor Econometrics and Operations Research

Finite-Sample Bias Corrections for Local Projections: Post-Estimation Corrections vs. Long-Difference Estimation

Primo's elevator pitch
Macroeconomists and central bankers need credible estimates of the effects of an interest rate hike on inflation, unemployment, and growth. Local projections are a widely used method for estimating these impulse responses over time, but the standard local projection estimator is biased in finite samples. My thesis tackles this problem by introducing a new estimator designed to reduce that bias. I evaluate its performance with extensive Monte Carlo simulations and benchmark it against two recently proposed bias corrections in the literature. The results show that the choice of correction is economically meaningful, with the novel estimator being favourable for certain bias–variance preferences and supporting more robust inference.