Our approach to calculating the incidence of tariffs is strongly informed by Clausing and Lovely (2024a).
In particular, we adopt their reasoning as to why the incidence falls on consumers (a view shared almost universally among economists) and whether to account for changes in behavior. We also adopt their assumptions about the policy: a 20 percent tariff on goods from all countries except China for which the tariff would be 60 percent and the assumption that the 60 percent would be a leveling up on past tariffs not on top of tariffs previously imposed (Clausing and Lovely, 2024b).
Although, as described below, our methodology differs from that of other analysts, such as Clausing and Lovely (2024) and Mulholland and Duke (2024), we reach similar conclusions. Our levels appear to be slightly lower than Duke and a bit less regressive than Clausing and Lovely. Our levels are higher than Clausing and Lovely because they do not distribute the entire amount attributable to the U.S. household sector.
A distinction between the approach used here and other analyses is that we make use of the consumption tax module of the ITEP Microsimulation Tax Model. The contours of that module are discussed in the methodological appendix of Davis et al. (2024). We also use a more complicated path from import data to household incidence. Our starting point is the Bureau of Economic Analysis (BEA) Import Matrices in conjunction with the BEA Use Tables. From these tables we find direct imports purchased as well as allocating imported inputs through to these categories of final purchasers. The final purchasers are in 17 categories in the data reflecting purchases for personal consumption, non-residential private fixed investment (businesses), residential private fixed investment, the federal government and state and local governments.
To distinguish between imports from China and the rest of the world, commodity level trade data are used and aligned with the BEA Use Table commodities.
For the portion of the imports that go to personal consumption, the BEA PCE Bridge is used to facilitate linking to the consumption items in the ITEP consumption tax module. Tariffs are calculated and distributed by the model. This constitutes most of what is distributed.
For the other final purchasers, we adopt different approaches. We assume that no tariffs are paid on imports by the federal government—or if they are, they are payments to itself so they do not impact households. State and local governments are treated as passing the taxes on to taxpayers. Business purchases and imported inputs into exports are modeled consistently with the handling of business consumption taxes in the ITEP Model as described in the methodological appendix of Davis et al. (2024).
For the revenue estimate we use the elasticities employed by Clausing and Lovely (2024a) as well as their revenue offset.