Non-Tariff Measure

The terminology carries a judgment, and it is worth being explicit about which one is being made.

Non-tariff measure

A neutral, descriptive category: a policy other than a tariff that can affect trade. It carries no implication about purpose or legitimacy. A maximum residue limit is an NTM.

Non-tariff barrier

An evaluative claim: that a measure restricts trade beyond what its stated purpose requires, or that it is protection wearing a public-policy justification. This is a conclusion requiring evidence.

The distinction matters because the same measure can be both or neither depending on facts that a classification cannot capture. A country setting a residue limit on scientific evidence is regulating; a country setting one to exclude a competitor is protecting. The measure looks identical in a database. Determining which it is requires examining the evidence behind it, and the multilateral system provides forums for exactly that argument.

An international classification organises non-tariff measures into chapters by type, distinguishing broadly between technical measures — those specifying product characteristics or the processes behind them — and non-technical ones. The categories most relevant to agriculture recur across markets.

• Sanitary and phytosanitary measures — requirements protecting human, animal, and plant life or health: residue limits, contaminant limits, pest freedom, disease status, treatments, establishment approval
• Technical barriers to trade — technical regulations, standards, and conformity assessment: labelling, composition, packaging, quality specifications, and the procedures for demonstrating compliance
• Pre-shipment inspection and other formalities — requirements applied before or at export rather than at import
• Contingent trade-protective measures — anti-dumping, countervailing, and safeguard actions arising from investigations
• Quantity-control measures — non-automatic licensing, quotas, and prohibitions
• Price-control measures — arrangements affecting the price of imports
• Export-related measures — restrictions, taxes, and licensing applied by the exporting country

The conformity assessment element deserves separate mention because it is where cost often concentrates. A requirement may be perfectly reasonable in substance and still be onerous in demonstration: the question is not only what a product must satisfy but who must certify it, using which laboratory, recognised by whom, and at whose expense. A measure and the procedure for proving compliance with it are different things, and either can be the actual obstacle.

The multilateral system does not prohibit non-tariff measures — countries have every right to protect health and set technical requirements. It disciplines how they may be adopted and applied, which is a different and more workable ambition.

The SPS framework covers measures protecting human, animal, and plant health, and orients them around scientific justification, encouraging measures based on international standards while permitting higher levels of protection where these are justified. The TBT framework covers technical regulations, standards, and conformity assessment, and turns on avoiding unnecessary obstacles and on non-discrimination. Both carry transparency obligations — notification of proposed measures, enquiry points — so that trading partners can see what is coming and comment before it binds.

Members also raise specific trade concerns about each other’s measures in the relevant committees, which functions as a forum short of formal dispute settlement. This is where most disagreements about agricultural measures actually happen, and the fact that a measure has been raised there is a matter of record rather than a finding against it.

A tariff has a number. A non-tariff measure does not, and the desire to compare the two has produced the practice of estimating an ad valorem equivalent — the tariff that would have the same trade effect as the measure. This is a useful device and an inference, not an observation.

• The estimate depends on the model, the specification, and the counterfactual assumed
• Measures interact, so attributing an effect to one of several applying simultaneously is not straightforward
• A measure’s cost differs by exporter — a large firm with laboratories and a smallholder cooperative do not face the same burden
• A measure that blocks trade entirely leaves no trade data from which to estimate its effect
• The counting of measures is not a measure of restrictiveness: many measures may be less burdensome than one

The last point is the trap. Databases record the incidence of measures, and a count is easy to compute and easy to misread. A country with many notified measures may be transparent and well-regulated rather than restrictive, and a country with few may simply not notify. Incidence is not intensity, and neither is legitimacy.