Matt Yglesias among his Predictions for 2025 sees Trump doubling the average US tariff rate from around 2 ½% to 5%. I have already written about the effects that tariffs have on inflation (it depends on the Fed) here. In this post I want to turn to look at the negative efficiency effects of the tariffs that Yglesias predicts.
As usual when you ask an economist a question, the answer is “it depends.” But with Matt’s prediction it matters more than most because it is really important how we get to the change in the average. The damage a tariff does – the amount by which it causes transactions to take place at both a changed price and as changed quantity -- increases with the square of the tariff rate.[1] Change” equals damage” That means that if the average were achieved with an additional 2½% on all imports which would do, say, $100 billion in damage, the average if achieved by 5% more being applied to only 50% of imports, the damage would be $400billion. If the additional tariffs were applied to only 8.3% pf imports at the 30% rate that Trump has mentioned for Canada and Mexico, the damage would $1440 billion. The same average tariff can do more of less damage depending on how uniformly it is applied.
Technically this arithmetic applies only _micro_-economically. Taking account of the _macro_-economic effect of tariffs, “strengthening” the exchange rate, reduces the estimated damage to users of the tariffed goods but recognizes the damage to exporters. The disproportionate effect of higher, more concentrated rates persists. Another specificity of this analysis is to assume that the tariff revenue does not affect macroeconomic demand, the fiscal deficit.[2] A change in the deficit would be an additional macroeconomic effect.
The absolute amount of damage depends not only on the rate of the tariff but on how much transactions change in response to the tariff. Tariffs on goods whose supply or demand is not reactive (low “elasticity” of demand or supply) cannot do much damage; they only redistribute income and raise revenue.[3] These elasticities depend not only on the nature of the tariffed goods themselves (and exports, let’s not forget), but on how easily producers and users can substitute other similar goods for the tariffed items.
The discussion so far has been carried out along traditional lines of tariffs only on final goods. Here, for example, is Peterson Institute for International Economics on the disadvantages for the music industry of tariffs directed at China’s exports of musical instruments. It points out both the real costs, the shift to higher cost imports, as well as the income distribution from purchasers of low-end musical instruments.
The points are all well taken but over look how much _worse_ tariffs are when applied to goods that are inputs into the production of other goods. Some analyses (Jon Murry in “Concerns About Tariffs/ for example) do point out that the effects of tariffs are different when some of the tariffed items are intermediate goods (steel for example) that consumers do not purchase but which will still impact prices of goods made from the tariffed input. The emphasis, however is that the cost to consumers extends beyond directly tariffed goods.
These are all efficiency effects on users of tariffed and tariff ridden goods as they shift into less preferred patterns of use. There is an effect on the production side, however, that is seldom considered but which that becomes important when tariffs are applied unequally on intermediate and final goods. For this we need to look not just at the “nominal” protection afforded by tariffs by raising the price of the final good but at the “effective” by raising the returns to the factors of production.
Let’s take a simplified example: stainless steel cookware. Suppose that in the absence of tariffs the cost of the product were $100 made up of $50 in steel and $50 of profit and wages. What happens if there is a 30% tariff on both cookware and steel
No Tariff On Both Only Cookware Only Steel
Cookware 100 130 130 100
Steel 50 65 50 65
Wages and profit 50 65 80 35
The revenue from the cookware rises by 30% and cost of steel rises by 30% so the wages and profits can also rise by 30%. The “effective protection”, the returns to factors of production is the same 30% as the “nominal” protection of the final good. If the tariff is applied on just the cookware, however, returns available to workers and owners rises to 80, an increase of, effective protection of, 60%. In contrast, if the tariff is applied only to steel (suppose cookware were exported) the income available for wages and profit falls by 30%, it has an effective protection of negative 30%
How changes in effective protection distributed between just higher (or lower) incomes for extant workers and owners and how much of additional resources will be drawn in or pushed out will vary by product line but the combination of non-uniform tariffs and intermediate goods can create extreme incentives and disincentives. And the efficiency effect still rise with the square of the rate.
But economic inefficiency is just chapter 1. To manage this kind of system (to avoid the most absurd results that can result) requires a huge bureaucracy (“Elon Musk, call your office!”) and invites varying degrees of legal and illegal corruption in its administration.
Image prompt: Merchants haggling over a transaction with a smirking tax collector with outstretched hand standing by.
[Standard bleg: Although my style is know-it-all-ism, I do sometime entertain the thought that, here and there, I might be mistaken on some minor detail. I would welcome comments on these views.]
[1] Change” equals “damage” because we assume that the transacting parties have already found the price and quantity that is to their mutual maximum benefit. Cases where tariffs can achieve an economy-wide benefit beyond the scope of the transacting parties, where tariffs can actually do more good than harm are discussed here.
[2] This is the standard way economist analyze changes in microeconomic policies and in the present context probably politically the case. In the “extension” of the 2017 “Tax Cuts for the Rich and Deficits Act,” income taxes will be reduced as much as Congress dares increase the deficit.
[3] Of course, if the objective were to raise revenue while minimizing damage, a VAT would be called for. If the objective were to favor certain lines of production (“industrial policy”) then subsidies rather than tariffs would be preferred as they would encourage production for export not just the domestic market. And other instruments are available to redistribute income. The general antipathy that economist have for tariffs s that they are almost never the right instrument for even a worthy objective.
A good analysis. I will just mention the late Max Corden who developed the concept of effective protection, which is essential to an understanding of these issues. Worthy of a Nobel award, but work done in Australia is always under-rated in the Northern Hemisphere (see also, Trevor Swan).