Where Do the Data Come From? Endogenous Classification in Administrative Data

11.11.2024 14:15 – 15:30

GTDW@IEE

Abstract:

Classication systems group heterogeneous objects such as products into categories called codes and dene what level of policy, such as taris, will be applied to each group. I develop a theory of endogenous classication and show that accounting for the motives underlying the design of classication systems has important implications for empirical research. I build a model in which the choice of which things to group together reects the objectives of a policymaker. A ner classication better targets policy but is harder to design and enforce. The degree to which heterogeneous objects are grouped together will vary across codes within a classication system and may be systematically related to policies and attributes of interest to both the econometrician and the policymaker. Taking the classication of U.S. imports as a leading example, I show that its design is consistent with this theory. Product codes vary greatly in their size and specicity codes are more subdivided when taris are high, and when codes are more subdivided goods within them are more similar. Heterogeneity in product attributes within codes leads to large biases when estimating parameters such as demand and supply elasticities. I show that the amount of heterogeneity and therefore the size of this bias is correlated with taris as expected under a theory of endogenous classication, and correcting it changes our understanding of important empirical relationships such as the one between taris and demand elasticities.

Bioography :

I am an assistant professor of economics at Dartmouth College. My research is in international trade and political economy.

Lieu

Bâtiment: Uni Mail

Boulevard du Pont-d'Arve 40
1205 Geneva
Room M 3250, 3rd floor

entrée libre