The U.S.-China trade war represents a natural experiment in the sense that we have not seen such wide-ranging increases in tariffs since the 1930s, when Congress passed the Smoot-Hawley Tariff Act (Bown and Zhang, 2019). Not surprisingly, applied trade economists have already conducted in-depth research on the impact of the trade war so far, the most notable being Amiti, Redding, and Weinstein (2019), Alberto Cavallo et al. (2019), and Pablo Fajgelbaum et al. (2020). This blogpost is a summary of the key results reported in these studies.
Of the studies cited, Fajgelbaum et al. is perhaps the most detailed. The authors constructed a monthly panel data set using publicly available tariff schedules issued by the U.S. International Trade Commission (USITC) along with U.S. import and export data published by the U.S. Census Bureau, where the tariff data are defined at the eight-digit level of the Harmonized System (HS), and the import data are defined at the HS-10 level. In addition, data on retaliatory tariffs were collected from the Ministry of Finance for China, the Department of Finance of Canada, the Office of the President of Mexico, and the WTO (covering the EU, Russia, and Turkey), tariffs being measured at the HS-6 level.
During 2018, U.S. tariffs were targeted at 12,043 specific products at the HS-10 level, where in 2017, these imports were valued at $303 billion, accounting for 12.7 percent of total U.S. imports. The average ad valorem tariff increased by from 2.6 to 16.6 percent. In terms of retaliatory tariffs on U.S. exports by Canada, China, Mexico, Russia, Turkey, and the EU, these accounted for $127 billion of U.S. exports, 8.2 percent of total exports, covering 8,073 products.
It is very clear from the data that the U.S. tariffs were mostly targeted at China, and Chinese retaliatory tariffs against the U.S. dominate, supporting the contention that the trade war has essentially been between these two countries. In 2018, the U.S. targeted 11,207 products accounting for 49 percent of total imports from China, tariffs increasing on average from 3.0 to 15.5 percent, while China targeted 7,474 products, tariffs increasing on average from 8.4 to 18.9 percent. The data also show that the most protected U.S. sectors were primary metals, machinery, computer products, and electrical equipment and appliances, while U.S. trading partners targeted different products, most notably agricultural imports.
Given their monthly panel data set, Fajgelbaum et al. undertook a detailed empirical analysis of the effects of the trade war on the U.S. economy. They first conducted an “event” study which consisted of comparing targeted and non-targeted U.S. imports and exports. In the case of imports, they found that their value and quantity declined by 20 and 23 percent, respectively. Importantly, they also present initial evidence that the incidence of U.S. import tariffs was borne entirely by U.S. consumers, tariff-inclusive unit values of imports increasing significantly as compared to before-tariff unit values which did not change. These authors also found a similar pattern in the case of exports, where their value and quantity fell by 24 and 25 percent respectively, with no change in their before-tariff unit values, i.e., there was, complete passthrough of retaliatory tariffs to foreign consumers.
Fajgelbaum et al. also used econometric methods to evaluate the impact of tariff increases on U.S. import demand and foreign export supply. Their results, which are statistically significant, show that both the value and quantity of U.S. imports declined in response to the application of tariffs, Amiti, Redding, and Weinstein finding similar effects in their study. However, Fajgelbaum et al. also found that there was no impact of U.S. tariffs on before-tariff unit values. The latter result provides further support for the argument that there was complete passthrough of the tariffs to tariff-inclusive prices borne by U.S. consumers. Similar results are reported for the impact of retaliatory tariffs on U.S. exports – there were significant declines in both the value and quantity of exports, but there was no reduction in before-tariff unit values by U.S. exporters. By contrast, Cavallo et al. (2019) found that there was imperfect passthrough of these tariffs to Chinese import prices of agricultural products.
The finding that incidence of U.S. tariffs has been almost entirely borne by U.S. consumers is consistent with the results of other studies using different estimation methodologies. It is also a surprising result given the importance placed on the terms-of-trade argument in the international economic analysis of optimal tariffs, as well as the empirical literature that has found less than complete passthrough of exchange rate shocks (Goldberg and Knetter, 1997). Over a longer time period, it might be expected that exporters would eventually cut before-tariff prices, especially if there was resolution of exporter uncertainty about how long the tariffs will remain in place. Interestingly, a follow-up study with data for 2019, shows that there has been some variation across sectors, e.g., U.S. tariffs led foreign steel exporters to lower their before-tariff prices (Amiti, Redding, and Weinstein, 2020).
The final step in the analysis of Fajgelbaum et al. was to quantify the effects of the trade war in 2018 using a computable general equilibrium model of the U.S. economy calibrated at the county level. Their results were as follows: first, U.S. consumers of imported goods in aggregate lost $51 billion due to higher prices; second, U.S. exporters saw an increase in their income of $9.4 billion; and, third, U.S. tariff revenue totaled $34.3 billion. Therefore, the net effect of the trade war was an aggregate loss of U.S. real income of $7.3 billion, which can be thought of as an approximation of deadweight loss. The latter number compares to Amiti, Redding, and Weinstein’s estimated net real income loss of $8.2 billion. As is often the case in applied trade analysis, the net economic effects are relatively small, but the re-distributional impact of tariffs on consumers have been substantial. Recent research suggests that subsequently this had a significant impact on consumption behavior, measured by reductions in county-level automobile sales (Waugh, 2019).
Importantly, contrary to what the administration has claimed, the results reported in Fajgelbaum et al., and in similar studies, clearly shows that the incidence of import tariffs implemented in 2018 has been entirely borne by U.S. consumers, any terms-of-trade effects on the import side being insignificant. Also, if there had been no retaliation by China and other countries, there would have been a modest U.S. real income gain of $0.5 billion in 2018 due to significant terms-of-trade effects on the export side. In other words, the logic of power-based bargaining only ever had the potential to work if China had not retaliated against the increase in U.S. tariffs.
In summary, the published empirical research on the U.S.-China trade war suggests that the administration’s choice of power-based bargaining over trade policy has been rather short-sighted. Unilateral imposition of tariffs against China and other countries has inflicted non-trivial economic damage on the U.S. economy, and runs the significant risk of undermining the multilateral trading system established in the post-war period under the General Agreement on Tariffs and Trade (GATT) and its successor the World Trade Organization (WTO).
Note:
This blogpost draws on a recently completed paper co-authored with Professor Daniel Chow (Ohio State University Moritz College of Law): “Understanding the Economic and Politcial Effects of Trump’s China Tariffs”.