Or, when 10% is not just 10%. (GM edition)

This is a reprint of a 2018 post. It takes on a heightened importance when one considers (1) the fact that tariffs on Canadian imports will hit a lot of inputs, rather than final goods, and (2) the integrated nature of the American auto industry.

An effective rate of protection (ERP) calculation takes into account the fact that domestic value added might be less than total value added – i.e., there is substantial imported value added in the final good. Suppose a 10% tariff is applied to motorcycles, so that a totally US made motorcycle which originally sold for $25000 (under free trade) can now sell for $27500 because of the protection from foreign imports. In this calculation, effective protection equals nominal, i.e., 10%. However, suppose imported inputs used in motorcycles were $15000. Then the effective rate of protection is 25% (=((27500-15000)/10000))-1).

Perhaps even more important, suppose instead a 10% tariff is applied to the imported inputs into motorcycle. Then the effective rate of protection is -15% (=((25000-16500)/10000)-1). To redress this negative rate of protection, the protection on inputs would have to be coupled with a 6% nominal tariff on motorcycles. The more pervasive vertical specialization (i.e., more specifically, the greater the share of imported components), the more prominent this problem.

And that’s a final good tariff just to even out things domestically. If you want to export the motorcycles, well, you’ll need an export subsidy. Under current WTO agreements, those are generally illegal.

Some things to consider as we go down the path of protection — and protection that is unlikely to reduce the trade deficit [1]

For a discussion on ERP from an expert, see Doug Irwin on pages 156-157 of International Economics (Cambridge University Press, 2025) .