Jour Fixe - Household Surveys and Machine Learning

Heterogeneity and the MPC: Observations from Machine Learning

Using a special Dutch survey that collects self-reported answers by households (HH) to unexpected hypothetical positive/negative income shocks, we use machine learning (ML) methods to try and identify distinct groups/clusters of HH’s that respond differently to the MPC. Our goal is two fold:

1.       Compare our findings with that of traditional least squares methods.

2.       Highlight the potential of ML methods in the HH finance/macroeconomics literature