Measuring trade in value added with firm-level data

Global Value Chains have proliferated economic policy debates. Yet a key concept - trade in value added - is likely mismeasured because of sectoral aggregation bias stemming from reliance on inputoutput tables. This paper uses comprehensive firm-level data on both domestic and international transactions to study this bias. We find that sectoral aggregation leads to overstated trade in value added and, correspondingly, understated import content of gross exports. The economic magnitude of the estimated bias varies from moderate to large - at 2-5 p.p. of gross exports for Belgium and 17 p.p. for China. We study how the interplay between within-sector heterogeneities in firm import and export intensities and firm size determine the magnitude of the sectoral aggregation bias.