Trump's Canada tariff hangs $10 billion a year on Section 338's first outing

Half of $20 billion is $10 billion, which is the run-rate duty this order creates if Canadian shipments hold at their pre-tariff level ^[2][14], about $27 million for each day it stands ^[15] and something near $190 million banked in its first week ^[16]. Read that as a ceiling rather than a forecast. A 50% wedge suppresses the volume it taxes, and the covered goods include cement and hockey sticks ^[9], freight substitutable enough to stop crossing a border once the landed price moves by half.

The attraction of a 96-year-old clause is procedural. The trade statutes Congress passed afterwards hand the president tariff power but confine it to defined circumstances, national security threats and foreign currency crises among them, and they oblige the government to investigate and clear other requirements first ^[10]; Section 338 as invoked arrived with no calculation of the dollar damage at all ^[9]. That is the speed the administration bought, and it is also the exposure it now carries, because Peter Harrell and Jennifer Hillman of Georgetown read the section to permit only tariffs that offset measured harm ^[8]. (Trump's own reading of Smoot-Hawley is that its levies came too late to rescue the American economy ^[13].)

Two respectable positions sit inside the same reporting. Ryan Majerus of King & Spalding, a former US trade official, calls the law a blank canvas precisely because nobody has ever litigated it ^[3]; Sara Albrecht, whose Liberty Justice Center already carried an earlier set of Trump tariffs to a win at the Supreme Court, says the strong argument is that the 1962 and 1974 trade acts superseded it ^[18][11][17]. This is probably wrong, but I would price supersession below the offset problem: implied repeal asks a court to find that Congress quietly killed a provision it never amended, while the offset defect is already on the face of the record, in a missing dollar figure and in a target list reaching goods with no connection to the dairy, auto and beverage complaints the order was built on ^[6][9].

The administration has skipped the investigative work those later statutes require ^[10], which is both the return on reaching for Section 338 and the reason a loss is survivable: void the order and much of the same duty can be rebuilt under the 1962 or 1974 authorities ^[17] with the paperwork attached, months later and probably narrower. The thing that would show I have this backwards is a published injury calculation tying the statute's 50% ceiling ^[4] to a measured loss for American dairy and auto exporters, alongside alcohol exporters ^[6]. Put a credible number on the harm and the offset argument stops doing any work.