Working Paper: NBER ID: w3059
Authors: Harry P. Bowen; Leo Sveikauskas
Abstract: Recent theoretical developments have cast doubt on the reliability of the commonly used cross-industry regression as a method for inferring a country's abundant factors. This paper examines the empirical importance of these theoretical cautions by comparing regression derived estimates of factor abundance with both revealed and actual factor abundances for thirty-five countries and up to twelve resources. Trade imbalances are found to importantly affect the regression estimates and we therefore derive and implement a theoretically consistent trade balance correction. The results indicate that despite theoretical concerns, the regression measures are often reliable indicators of revealed factor abundances. The results therefore enhance the credibility of the findings of the numerous regression studies that have been conducted over the past thirty years.
Keywords: factor abundance; trade imbalances; regression analysis
JEL Codes: No JEL codes provided
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| trade imbalances (F14) | regression estimates of factor abundance (C29) |