Incorrect Nesting Structures in Models Lead to Biased Welfare Predictions.
The article examines how nested logit models perform when there are errors in their structure or underlying process. The researchers found that choosing the wrong nesting structure or using an inconsistent stochastic process can lead to biased welfare predictions. They used Monte Carlo experiments and analytical results to show the impact of these errors on welfare estimates. The study also looked at different criteria for selecting nesting structures.
