In theory, retaliatory tariffs probably aren't first step in the WTO dispute resolution process.
Instead, the dispute is supposed to go through consultations, which take some 6.6-months on average, followed by a panel decision if necessary, which can take an additional 15.1 months etc. (Approximately two thirds of WTO-registered disputes were actually solved in the consultations stage.) Following an arbitration,
If defendant stays idle, or if a compliance panel (and/or the AB [Appellate Body], as the case may be) pronounce on the inadequacy of adopted measures, then defendant might be facing retaliation. The framers of the DSU did not manage to agree on a clear method to calculate the amount of retaliation. They used the term ‘substantially equivalent concessions’ that would serve as the legal benchmark for calculating retaliation. Disputing parties might disagree on the amount of retaliation, and disputes on this score will be submitted to an Arbitrator (the original panel, whenever possible), which will decide on the appropriate amounts (Article 22.6 of DSU). Awards by the Arbitrator on this score are not subject to appeal. Retaliation can lawfully take place only following authorization to suspend (tariff) concessions (that is, raise the level of customs duties on imports from the recalcitrant WTO member). The defendant is still under the obligation to eventually bring its measures into compliance, and has to observe specific reporting requirements to this effect (Articles 21.6 and 22.8 of DSU). Suspension of concessions is ‘temporary’ as per the explicit wording of the DSU (Article 22.8 of DSU), but no statutory deadline for its expiry is provided. Consequently, whereas de jure the DSU (Article 22.1) calls for ‘property rules’, that is, for the obligation to perform the contract, de facto it tolerates ‘liability rules’, that is, the possibility for authors of an illegality to ‘buy their way out of the contract’.
So, it looks like (in theory) retaliation in the WTO framework can only legally happen after arbitration. On the other hand, there's was no shortage of retaliatory tariffs worldwide, recently. Was this always the case for trade disputes ever since WTO or GATT came about, i.e. in practice, does retaliation often work in parallel with the WTO dispute resolution? (Some statistics on this would be more interesting to me than mere opinions.)