Intermediate inputs, external rebalancing and relative price adjustment

The traditional multi-sector macro model without production inputs is a value-added trade model. This paper shows that calibrating such a misspecified value-added trade model to available gross-flow trade data - a common practice in the literature - can lead to mismeasured (i) preference weights and (ii) price elasticities. Further, the calibrated model can give substantially different predictions regarding the relative price response to external rebalancing, when compared to a preferred alternative model with inputs that is consistent with gross-flow trade data. We find that mismeasured preference weights and price elasticities both contribute sizably to deviations in model predictions and estimate correctly-measured parameters for the value-added trade model.