BLOG POSTS ARE PRELIMINARY MATERIALS CIRCULATED TO STIMULATE DISCUSSION AND CRITICAL COMMENT. THE VIEWS EXPRESSED ARE THOSE OF THE INDIVIDUAL AUTHORS. WHILE BLOG POSTS BENEFIT FROM ACTIVE UHERO DISCUSSION, THEY HAVE NOT UNDERGONE FORMAL ACADEMIC PEER REVIEW.
James Roumasset*
“Give me a one-handed economist. All my economists say, ‘on the one hand… on the other.’”
President Harry Truman
Do tariffs increase or decrease economic welfare? It depends on the model.
Aggregate Supply and Demand
Some economists, e.g. Kevin Hassett and Peter Navarro, try to infer tariff consequences from the aggregate demand (AD) equation, Consumption (C) + Investment (I) + Government Expenditures (G) + (Exports X — Imports M), or C + I + G + (X – M). But noting that C, I, and G are composed of both domestically produced (d) and import components (m), we can rewrite AD as Cd + Cm + Id + Im + Gd + Gm + (X – M). Netting out imports, we have Cd + Id + Gd + Gm + X, which is aggregate supply (AS) by definition. That is, AD = AS is an identity (represented by the circular flow diagram of gross domestic product) and useless by itself to explain policy consequences.
Keynesian economics
In simple Keynesian economics, we convert the identity into a theory by making AD exogenous and articulating things that make it go up or down. In this model, imports drive prices up and AD down, lowering the point at which AS = AD. But it’s all rather vague, lacking a theory of how much prices go up and how much they reduce AD.
Foreign exchange
Assuming perfect competition, no externalities, and no retaliation, we can analyze the effects of a tariff by means of the supply and demand for foreign exchange. A tariff on imports shifts the demand down, lowering the price of foreign exchange (increasing the exchange rate). This causes a movement down the supply curve of foreign exchange, lowering exports. Both imports and exports fall, lowering the gains from trade. The diagram shows the amount of welfare loss as an excess-burden triangle.
Of course, the assumptions are severe. Stiglitz and Greenwald (2006) showed, for example, that tariffs can boost economic welfare by correcting a positive dynamic pecuniary externality associated with interdependent investments in manufacturing. This provides a case for selective tariffs as part of a country’s industrial policy. Of course, the government has to know where the important interdependencies are.
Static noncooperative games
Assuming away the possibility of bargaining, we can analyze the effect of US tariffs as a rightward shift in the US reaction function, where US tariffs are on the horizontal axis and ROW (rest of world) tariffs are on the vertical axis. The noncooperative equilibrium tariffs increase. This is the implicit model underlying Larry Summers’s critique of Trump tariffs. He implicitly assumes that tariffs beget retaliatory tariffs.
It is sometimes argued, in effect, that the US is not currently on their rational reaction function in the first place, i.e. that they have failed to respond to “unfair” policies of their trading partners. In this view belated retaliation begets at least short-term gains.[1]
Dynamic noncooperative games
In a model of repeated noncooperative games, a strategy of tit-for-tat reciprocal tariffs can lead to lower tariffs (cooperation), but the required assumptions (patience and far-sightedness) are rather stringent. (See also discussion here).
Cooperative games
Allowing bargaining plausibly increases the prospects for reciprocal tariffs leading to lower tariffs (e.g. Nash Bargaining Solution, Shapley Value, or Harsanyi Dividend) and higher economic welfare. If the country with lower tariffs imposes reciprocity, the other country revises their subjective payoffs to noncooperation downward thereby improving cooperation. While this outcome is plausible, cooperative game theory and experimental economics suggest that other outcomes are also plausible.
Supply Shocks
In the early days of Covid, there was talk of how its initial shock would cause “scarring,”[2] of supply chains, i.e. that supply shocks may leave lasting effects. The same problem may arise if tariffs increase. To the extent that economies are already integrated, e.g. such that auto production involves parts being sent sequentially back and forth across country boundaries as intermediate products are assembled, it may be difficult to rebuild supply chains.
Concluding Remarks
Two apparently contradictory things can be true at the same time. On the one hand, there is the much-ballyhooed uncertainty regarding current U.S. tariff policy. Not only has the on-again-off-again nature of tariff threats created uncertainty, but we don’t know to what extent tariffs will tend to equilibrate at higher or lower levels, and we don’t know what the longer-term effects will be (including effects on exchange rates and Fed actions). On the other hand, said uncertainty may not have large economic consequences. Jay Powell was prudent to say recently (March 19, 2025) that the Federal Open Market Committee doesn’t know what role tariff uncertainty has played in the increased short-term inflation expectations. He also said that it is longer-term inflation expectations and the 10-year Treasury bond rate that matter more for the investment outlook and that these have not increased.
Powell’s modesty is exemplary. Economists would do well to condition their remarks on particular models and assumptions.
References
El-Erian, Mohamed. 2025. What insights does Game Theory offer on the U.S. tariff policy approach? https://x.com/elerianm/status/1885714207246627300
Greenwald, B. and J. Stiglitz. 2006. Helping Infant Economies Grow: Foundations of Trade Policies for Developing Countries. American Economic Review 96 (2): 141–146.
* Professor Emeritus, UHERO and Department of Economics. Thanks to Lee Endress, Katya Sherstyuck, Kimberly Burnett, Chris Wada, and Victoria Rhinebolt for helpful suggestions.
1 El-Erian (2025).
2 See e.g. comments in 2020 by Mohamed El-Erian and Janet Yellen.