How do permanent tariff increases affect inflation and unemployment?
Core argument: Using a yearly VAR with 2 lags on tariff changes, unemployment, and CPI inflation, the authors estimate that over 1869–1939 a roughly 4 percentage point permanent tariff increase lowered inflation by 2 points on impact and raised unemployment by about 1 point.
To estimate the dynamic causal effects of tariffs, we use a yearly VAR (vector autoregression) with 2 lags and (at least) 3 variables: tariff changes, unemployment, and CPI inflation. A careful review of the major changes in US tariff policy since 1870 shows no systematic relation between the state of the cycle and the direction of the tariff changes, as partisan differences on the effects and desirability of tariffs led to opposite policy responses to similar economic conditions. [Thus] Figure 5 plots the “OLS estimates”, i.e., treating innovations to the average tariff rate as independent of the business cycle. The results are remarkably consistent across sample periods; whether we use the full historical dataset (1869-2020), only pre-World War II data (1869-1939), or only post-World War II data (1946-2020). We find that a permanent tariff increase has a large effect on inflation and unemployment. Over the pre-WWII sample, a roughly 4ppt permanent increase in tariff rate lowers inflation by 2ppt on impact and raises unemployment by about 1ppt. The estimates are economically large and strongly statistically significant. Over the post-WWII sample period, the estimates are more uncertain given the much more limited variation in tariff rate over that period, but show the same conclusions: higher tariffs lower inflation and raise unemployment.




