Trade costs, home bias and the unequal gains from trade

In a recent paper, Fajgelbaum and Khandelwal (2016) develop a methodology to quantify the distributional consequences of trade. Their approach relies on aggregate expenditure data and on a non-homothetic gravity equation based on an Almost Ideal Demand System (AIDS). In this setup, we show that the structural parameters governing the welfare gains are highly sensitive to the determinants of spending on domestic goods. We extend their model by allowing for a home bias in tastes or, alternatively, for more complex trade costs. While Fajgelbaum and Khandelwal (2016) find that trade typically decreases the relative price of the goods consumed by poor households, we show that the pro-poor bias of trade becomes weaker or can even turn slightly pro-rich for most countries depending on the specification.