Frank Walsh: This paper makes a number of important contributions. In terms of data, a new worker firm panel is assembled by combining data across a range of surveys. This allows the author to control for a range of characteristics including firm-level fixed effects. This makes the evidence much more convincing. Although it is unfortunate that data constraints prevent the author from tracing workers over time, a range of worker individual-level controls are included.
Hummels et al. (2014) is discussed as a reference for the instrumental variables (IV) approach adopted in the paper. In this approach, the world export supply net of home country imports is used as an instrument for offshoring. Interestingly, Hummels et al. (2014) also use world import demand net of home country exports as an instrument for exports, which might be interesting for future work. The estimated IV effects for annual income...