New evidence on preference utilization

We analyse the degree of preference utilization in four major importing countries (Australia, Canada, EU and US) and provide evidence that preferences are more widely used than previously thought. For Australia and Canada, we have obtained a new dataset on imports by preferential regime that has so far not been publicly available. For the EU and US, we make use of more disaggregated data than previously used in the literature. We empirically test what determines utilization rates. In line with previous studies, we find that utilization increases with both the preferential margin and the volume of exports, suggesting that using preferences can be costly. However, we also find that utilization rates are often very high, even for very small preferential margins and/or very small trade flows, which contradicts numerous estimates that average compliance costs are as high as 2-6%. We extend the existing literature in relation to both data and methodological issues. In particular, we construct "pseudo transaction-level" data that allows us to assess more precisely when available preferences are utilized. Using this methodology, we obtain a more realistic estimate of what determines utilization. Rather than constituting a percentage share of the trade value, our findings indicate that utilization costs involve an important fixed cost element. We provide estimates for such fixed costs, which appear to be in the range of USD 14 to USD 1,500. -- Trade policy ; preference utilization ; fixed costs ; transaction data