WEBINAR RECORDING · 2022-04
Trade Costs and Trade Flows
In this conversation: Dennis Novy; discussant Yuan Zi; moderator Ben Shepherd.
Dennis Novy shows why a common trade-cost change can produce very different percentage trade responses across partners. Translog gravity makes the response depend on initial import shares. Yuan Zi probes whether this reflects heterogeneous elasticities, heterogeneous policy shocks, or model assumptions—and what follows for welfare.
Session information draws on the recording and its original description. Affiliations and policy context refer to the time of the webinar.
FROM THE CONVERSATION
Three takeaways.
What they mean in practice.
Prepared from the session recording using automatic transcription and editorial review. Follow the timestamp links for the original wording and context.
An average effect conceals bilateral variation
Watch 08:35 ↗Novy presents joint research with Natalie Chen in which trade-cost reductions produce larger percentage increases in initially thin trading relationships and smaller increases in heavily integrated ones. Currency unions illustrate the mechanism; the argument is broader than that application. A large percentage response in a small relationship need not imply a large increase in trade value.
Policy interpretation: Report partner-specific responses and distinguish percentage changes from absolute trade creation when evaluating integration policies.
The elasticity is a structural assumption worth testing
Watch 25:58 ↗The translog specification links the magnitude of the trade-cost elasticity inversely to import shares, instead of imposing a constant elasticity. Novy explains the estimation changes required and later acknowledges a trade-off: standard gravity fits aggregate trade extremely well, while this specification captures a dimension of heterogeneity that the constant-elasticity framework rules out.
Policy interpretation: Compare economically relevant predictions across specifications; a good overall fit does not establish that a model captures heterogeneous responses.
Trade-flow findings are not yet a welfare assessment
Watch 43:12 ↗Asked about welfare, Novy explicitly says the presented paper does not analyze it. He describes preliminary work suggesting heterogeneous welfare effects but emphasizes that it is not sufficiently developed. His Brexit comments are likewise predictions from the framework, not an estimated Brexit application.
Policy interpretation: Keep estimated trade effects, model-based extrapolations, and completed welfare calculations separate in policy briefings.
From the discussion
Yuan Zi asks how to separate heterogeneous policy-induced cost reductions from heterogeneous elasticities (36:56–38:24). Shepherd presses on model comparison (45:29–49:27) and causal inference (49:53–53:00); Novy acknowledges endogenous agreements and currency unions, while describing the paper's interpretation of that issue. Zi closes by urging researchers to make the mapping from reduced-form coefficients to structural interpretations explicit (57:26–60:05).
Recording references
- Recording: heterogeneous trade responses — 08:35–31:38
- Recording: discussant and audience exchange — 36:56–60:05
