Optimal Post-Hoc Theorizing

TL;DR

This paper formalizes the tradeoff between theorizing before or after empirical analysis in economics, arguing that in modern contexts, post-hoc theorizing is generally optimal due to the balance of learning effects.

Contribution

It introduces a Bayesian model to formalize the tradeoff and demonstrates that post-hoc theorizing is often optimal in the context of large datasets and mature theories.

Findings

Post-hoc theorizing is typically optimal in modern economics.

The model formalizes the tradeoff between theory-first and data-first approaches.

Empirical results support the advantage of post-hoc theorizing in large data environments.

Abstract

For many economic questions, the empirical results are not interesting unless they are strong. For these questions, theorizing before the results are known is not always optimal. Instead, the optimal sequencing of theory and empirics trades off a ``Darwinian Learning'' effect from theorizing first with a ``Statistical Learning'' effect from examining the data first. This short paper formalizes the tradeoff in a Bayesian model. In the modern era of mature economic theory and enormous datasets, I argue that post hoc theorizing is typically optimal.