The Non-MFN Effects of MFN Specific Tariffs

This paper explores the sources, extent and two consequences of the non-MFN effects of US MFN specific tariffs resulting from within-commodity cross-exporter variation in f.o.b prices. The first consequence is the bias against the developing countries exporting to the US, which is a result of these countries exporting low quality and low priced varieties (Schott, 2004). This bias is measured as the additional loss in imports faced by these countries, from US MFN specific tariffs as opposed to from a benchmark tariff vector defined as the average Advalorem Equivalent (AVE) of the MFN specific tariffs. For fixed world prices, my results using the Anderson-Neary Mercantilist Trade Restrictiveness Index (MTRI) suggest that doubling per capita GDP reduces AVEs by 14%§ . This is equivalent to an additional import loss of 5-8% or 60-80 million dollars, from the group of 120 developing countries exporting to the US. This value is as high as 15-24%, equivalent to 13-20 million dollars for a subsample of the 48 low income countries which export to the US. Another consequence of the Non-MFN effects of MFN specific tariffs is the additional deadweight losses (DWL) accruing to the US on account of US MFN specific tariffs. Since the DWL is proportional to the squares of tariffs, the AVEs of MFN specific tariffs being non-MFN, have an additional cross-country level of variation associated with them, inducing greater welfare losses. With fixed world prices, my results using the Anderson-Neary Trade Restrictiveness Index (TRI) suggest that levying MFN specific tariffs as opposed to the benchmark tariffs, increases the US trade restrictiveness and DWL by an additional 6% and 12% respectively. All my results are robust to five alternative specifications of AVEs, each of which control for the measurement and reporting errors in unit values to various degrees.