Measuring precise TFP growth under trade liberalisation: stripping away omitted price bias and demand shocks

We address two important sources of bias that remain unaccounted for in most analyses of firm-level productivity and trade policy; that is, omitted output price bias and demand shocks. Using firm-level data that includes product-level price and physical output measures, we mitigate the omitted output price bias typically encountered when using sectoral deflators. Besides being a primary determinant of firm survival, our precise controls for demand shocks also help account for quality differences between firms. We then test the methodology developed by De Loecker for use with firm revenue data and find that its accuracy depends on the precision of the price deflator. Results from Pakistan suggest that, after controlling for omitted price bias and quality differences through demand shocks, the impact of a 10% reduction in tariffs on firm-level productivity decreases from 1.26% to 0.36%. The net impact of trade liberalisation has been a mere 2.2%, with the biggest gains in the least protected segments. The estimates for precise demand shocks are reasonably close to the 3% improvement estimated using disaggregated price and output data with De Loecker et al. JEL Classification: F1, D22, D24, L23.