Shylock Holmes

Tariffs produce tax revenue which can be used to offset other taxes. The failure to include this in comparisons greatly slants the argument towards free trade in the eyes of economists, by conflating "lower distortionary taxes" with "lower tariffs specifically". 2/

Existing taxes, especially income tax, have huge deadweight losses - compliance is a big pain, distortions from lower incentives to work are damaging. Even if free trade is good, this justifies more agnosticism about a *revenue neutral tariff* than most economists display. 3/

Tariffs are much simpler and cheaper to collect. All you need is to control shipping, borders and air freight. The convoluted structure of income tax justifies an enormous surveillance apparatus and random punitive auditing that was used to target conservatives in the past. 4/

A tariff is a flat % tax, so helps flatten the tax structure. Many Americans pay no income tax, so have incentives to vote for increases. The US has one of the world's most progressive tax systems. Tariffs reduce this, and are more politically palatable than "tax the poor". 5/

A tariff is like a sales tax, with a rebate for domestic employment. Sales taxes and VAT taxes are generally not considered terrible as far as taxes go. Employing domestic labor has genuine positive local externalities relative to overseas labor. 6/

An employed person will have more friends, more life purpose, less likelihood of boredom and ennui that has driven the epidemic of opioids, alcoholism and suicide. Standard trade models ignore these externalities entirely. 7/

Put in alternatively, most trade models assume that leisure is a good, and labor is a bad. This is a great assumption when going from a 14 hr work day to 6 hrs, but very dubious when going from 6 hrs to permanent unemployment, *even if consumption is unchanged*. 8/

This same reservation would also apply to other labor-saving improvements like automation and AI. If labor enters non-monotonically in the utility function, lots of economic conclusions that were previously obvious are no longer obvious. 9/

The assessment that America's primary source of unhappiness is lack of consumption, as opposed to social ills from mass idleness, is similarly unclear. 10/

The benefit of such reductions in trade, automation etc. is not that they increase consumption by stimulating the domestic economy. They are predicted to reduce consumption. But they may, depending on quantities, increase welfare by increasing employment. 11/

All this depends greatly on the quantities. I is clearly possible to set taxes and restrictions too high. Is the externally of domestic production actually worth rebating 100% of a nationwide sales tax, as opposed to some smaller amount? Probably not. 12/

Is it possible to set tariff rates too high? Absolutely. None of this is a blank check. 13/

There are other more speculative things you could add. If you think that US dollar hegemony is a bad force in the world (or bad for inherently making US domestic industry uncompetitive), reducing international trade will reduce demand for dollars as hedging assets. 14/

The baseline assumption that exorbitant privilege is actually bad for the US is far from clear, but also not as silly as you might assume. 15/

I think it is possible to read all this and still think tariffs are on net a bad thing. But I find the case to be far more uncertain than most economists tend to assume, even when analyzed purely through the tools of economics. 16/

Postscript: There are also many other variants worth pondering that are probably better. If this were structured as a nationwide VAT, set at a higher rate for foreign goods to take into account positive externalities of domestic employment, would you maybe be in favor then? 17/

If so, we are more haggling over the specific implementation than the direct principle that trade can never be taxed. 18/

Finally, huge country differences are harder to justify under this model. The case for 25% on Canada, and 0% on most other countries, is especially difficult to make, in my opinion. 19/