The U.S. currently subjects imports to a dizzying schedule of tariffs and tariff exemptions. What is the most accurate way to summarize the overall level of these trade restrictions and their cumulative impact on American consumers?
One approach attempts to parse the array of announced tariff rates, weights them by each product’s share of imports, and reports the average. (Most prominent example: the “overall average effective tariff rate” from The Budget Lab at Yale.) Beyond the challenge of keeping up with announcements, revisions, and lags in implementation, this method requires assumptions to cover the many country-based and other exceptions that dilute the statutory tariffs in actual practice.
Another option bypasses these issues by going directly to the trade data, dividing the total tariff duties collected by the total value of imports. (See the “effective tariff rate” from the Penn Wharton Budget Model). As a measure of economic impact, however, this simplicity comes at a cost: It ignores the substitutions and spending cutbacks that consumers have already made in response to rising prices.
Minneapolis Fed Monetary Advisor Michael Waugh proposes an improved “trade restrictiveness index” (TRI) that is superior to prior approaches (Minneapolis Fed Staff Report 681, “How Restrictive Is U.S. Trade Policy?”). Waugh’s TRI translates the monthly reported trade data into a single number: the universal tariff rate that, if it were applied uniformly to every U.S. import, would cause the same economic pain as the actual, highly uneven tariff regime.
By this new measure, the U.S. economy experienced the equivalent of a 21 percent universal tariff rate in November 2025. This is twice the 10.5 percent effective tariff rate reported by the Penn Wharton Budget Model and about 24 percent higher than the average statutory rate calculated by The Budget Lab.
Waugh’s approach combines the strength of using actual trade data with a theoretical observation that other methods neglect: High tariffs hurt proportionately much more than low ones. In technical terms, the “deadweight loss” from distortions in economic activity does not increase linearly with tariff rates, but exponentially—with the square of the rate. The economic pain to consumers of a 10 percent tariff, for example, is four times the pain of a 5 percent tariff.
Measures that calculate an average tariff rate wash out the disproportionate harm caused by above-average tariffs. “A few very high tariff rates create more distortions than many moderate tariffs with the same average rate,” Waugh writes. The U.S. tariff structure by late 2025 featured a wide dispersion of implemented rates rising as high as 50 percent. Waugh’s TRI measure takes account of this heterogeneity and thus yields a much higher equivalent rate than a simple duties-to-imports calculation or another mean weighted tariff measure from the literature (Figure 1).1
Waugh’s index combines the findings of prior researchers for a near-real-time measure of current trade policy. The design of a trade restrictiveness index that expresses heterogeneous tariffs as a single, welfare-equivalent universal tariff traces to a 1996 article by Anderson and Neary. Papers by Feenstra (1995) and Kee et al. (2009) point to a mathematical shortcut that allows Waugh to quickly update his TRI using the latest Census Bureau trade data without the need for a dense general equilibrium economic model. Waugh intends to publish a regularly updated TRI via his Trade War Tracker website.
Beyond the top-line number, the TRI reveals notable trends across countries and industries. The TRI is relatively similar to average tariff measures in the case of China, where the applied tariff is high and fairly uniform across imports, with relatively few exemptions. For Canada and Mexico, however, the TRI is much higher than the popular measures would suggest (Figure 2).
Trade restrictiveness index versus alternate tariff measures
These differences arise because while most categories of traded goods are exempt from tariffs under the U.S.-Mexico-Canada Agreement (USMCA), some critical categories are subject to high rates—principally automobiles (25 percent) and steel and aluminum (50 percent). Canadian lumber also faces a high effective tariff rate (about 45 percent). The TRI reveals that despite the USMCA, overriding sectoral tariffs are significantly crimping trade with our neighbors. Brazil and Switzerland similarly have much more restricted trade than their average tariff rates would suggest.
On a sectoral level, Waugh notes that machinery and electrical equipment have elevated TRIs. These categories of goods come from many different countries, which have been hit with a range of “liberation day” tariff rates, and are characterized by many exemptions—all of which drive actual trade restrictiveness higher than an average tariff rate would suggest. The TRI measure similarly shows that pharmaceuticals, which look nearly duty-free under other measures, are actually subject to the equivalent of a 4 percent universal tariff.
Waugh notes possible limitations of the TRI measure, including that its efficient calculation depends on an assumption of common trade elasticity among all goods. Like other approaches that use tariff levels to express the cost of trade, it does not reflect other possible costs of tariffs, such as inflation or slower economic growth. However, as an easy-to-update measure that reflects the extra distortions created by a wide distribution of tariffs, the TRI better captures the economic burden of today’s new era of U.S. trade.
Read the Minneapolis Fed staff report: “How Restrictive Is U.S. Trade Policy?”
Endnote
1 This alternative mean weighted tariff measure is analogous to the mercantilist trade restrictiveness index from Anderson and Neary (2003), which calculates a uniform tariff that would result in the same level of aggregate imports. However, it does not incorporate the dispersion of tariffs as does Waugh’s TRI measure.
