People shopping in Midtown Manhattan in New York on Saturday, February 12, 2022.

Blog Name

Trade barrier reduction is the most important anti-inflation competition policy

Lawrence H. Summers (Harvard University)


Photo Credit: Sipa USA via Reuters Connect/Richard B. Levine

Edited excerpts of remarks at a PIIE event on March 30 discussing For inflation relief, the United States should look to trade liberalization, a Policy Brief by Gary Clyde Hufbauer, Megan Hogan, and Yilin Wang.


The Peterson Institute has again produced a very important study making a powerful, analytical, and statistical point with respect to a crucial public policy issue. This study reinforces my conviction that trade barrier reduction can while supporting consumer incomes and economic growth make a significant contribution to disinflation. This contribution dwarfs the potential impact of other competition policies that have preoccupied the White House.

As a discussant, I wish to talk about this paper at several levels.

First, is the core estimate that reasonable tariff reduction in the United States could take more than 1 percent off the CPI [Consumer Price Index] a reasonable one? My judgment after reviewing all the statistical machinations is that the answer to that question is decisively yes.

The key analytical point is this: When you reduce tariffs, you have two effects of reducing the price level. One is that people no longer have to pay the tariffs. That is the direct effect, and that is the effect that some previous Peterson Institute work had suggested was relatively limited. The other is that there are people who compete and firms who compete with the tariff good. And so just as if you reduce the price of Pepsi, you force Coca-Cola to reduce its prices, when you reduce the price of imported goods, you reduce the price of domestic goods as well. That turns out to be the larger of the two effects.

The indirect effect is, according to these estimates, on the order of two to three times the direct effect. And that's because competition is really an important thing. If you reduce the price of Fords, it has a big effect on GMs. But if you reduce the price of tennis rackets, it only has a small effect on the price of golf clubs. It depends upon how substitutable the goods are. These assumptions seem to me to be proven based on other work in other areas and to be plausible.

These estimates are, in a sense, conservative because they are estimates only of first round effects. When you reduce tariffs and reduce prices, it reduces wage demands. And when you reduce wage demands, that leads to subsequent reductions in prices and contributes to further disinflation. I think, prudently and appropriately, the authors have left those estimates out. But they could have been added as well. So,I think these should be thought of as reasonable, responsible estimates.

Additionally, these estimates meet a test that I always found very important as an economic policymaker, which is you can understand where they came from. It is not that we put all the numbers into the computer, and this is what came out. It is basic economic logic, carefully calculated.

The second question in evaluating this study is: Is the tariff program outlined a reasonable one? Or is this a program of extraordinary and implausible ambition? And I satisfied myself in reviewing the paper that it was a reasonable one in two ways. The first and more obvious of them is by applying my judgment to the list of specific measures through which it could be derived. The second was by looking at the figures in the paper on the magnitude of the Trump tariff increases. While it is probably unreasonable to think we're going to do something as big as a 100 percent reversal of the Trump increases, it is probably reasonable to think we're going to do something that's as big as reversing two thirds of the Trump increases, in part by reversing the Trump increases and in part in other ways. So I think this is a highly credible estimate and that it should be used as a new benchmark in discussions of the issue of the contribution of tariff reduction to disinflation.

A third question I asked myself is: How big a deal is the 1.3 percent decrease in the CPI that the authors estimate? The authors suggest that it's $800 per household. That is a lot of money for many households in this country. For a household that consumes 500 gallons of gasoline a year, that is the equivalent of more than a dollar a gallon increase in the price of gasoline.

Another way of thinking about the significance of tariff reduction is to compare it to other microeconomic measures that are discussed with regards to inflation. One should only use the term order of magnitude when one actually means a factor of 10, not just if one wants to say that something is important. This inflation reduction through tariff removal is in fact an order of magnitude more significant than anything that can be achieved over the next two to three years through antitrust action. It is probably two orders of magnitude more important than the meat packing fixation, which has been the subject of several presidential events.

I think the right estimate of the role of producer price manipulation in explaining high gasoline prices is zero. There are others who disagree. But the most extravagant claims for manipulation by oil companies with respect to gasoline prices would not rise to a tenth of what is at stake here.

So, in the sphere of microeconomic measures that might plausibly bear on inflation over the next two to three years, this is by far the biggest measure available, and it should be seen in that way. I hope the National Economic Council will take note that a competition policy agenda that does not include trade barrier reduction is not ultimately serious.

By the way, if you ask me for a second biggest measure over the next several years, it would be immigration policy, which has in common the Peterson Institute theme of promoting international integration and globalization. By the standards of those micro policy proposals to address inflation which have been emphasized, this is immensely important.

A fourth observation that I would make is that there are difficult conceptual questions about the role of microeconomic policies versus macroeconomic aggregate demand policies in the explanation of inflation phenomena. The effect of any change like this depends upon how monetary and fiscal policy vary. Ultimately what is vastly most important for inflation is macroeconomic policy and the manipulation of aggregate demand. Or to use two clichés from the economics literature: Inflation is about the price of goods in terms of paper and ultimately the quantity of paper versus the quantity of goods (with doctrinal debates about the role of the money stock versus the role of debt instruments and the like). So there is a primary role for macroeconomic policy.

It is also crucial to distinguish between high and rising prices, and that which causes high prices does not necessarily cause rising prices. That is why the authors of this study are careful to say this is a 1 percent-plus decrease in the price level, which then will take the edge off of inflation. They are careful to be conservative in their claims by not taking account of any spiraling impacts. By the way, they have actually done careful calculations of those spiraling effects, but they decided that the more prudent course was to be conservative in the estimates that they were providing.

What does all this say operationally for US trade policy? It does not say that for this reason the United States should immediately and forthwith remove all tariffs without getting anything in return from trading partners. Back in the arms control days, it was important to understand that by reducing your stock of weapons, you could get your counterpart to reduce their stock of weapons, too. But there is a big difference between having that insight and having the view that it's better to have as many weapons as you can, and that arms control is something to be avoided.

In the same way, the appropriate strategic objective of the United States with respect to trade policy needs to continue to be a more open world, not a more closed world. In addition to the anti-inflationary and pro-consumer benefits from a more open world, we will also get some measure of reciprocity if we are halfway skillful about it, which will serve American export industries and American competitiveness to a further degree.

Earlier Peterson work suggested gains from trade for US households exceeded a trillion dollars. I think the study that Peterson did some years ago, with Fred Bergsten and Gary Hufbauer having very major roles, calculating that it was over a trillion dollars in potential benefits that the United States got from trade in the form of lower prices, was a very, very important one. And I think we just need to keep banging the drum. One way of saying it is that real wages, the purchasing power of workers, is wages divided by prices. But you could increase that number either by raising the numerator or by lowering the denominator and they are both equally effective.

I think it is a shame that we have an office of the United States Trade Representative that so consistently reflects a range of special interest political constituencies rather than the general interest of American workers in lower prices. One of the things that President Clinton did well was when he spoke about trade, in addition to making the mercantilist arguments for trade, that it would generate more exports and so forth, he also made the David Ricardo argument for trade about the lower prices and higher standards of living that would result. This is why studies like this one are so important; they put these arguments into the air.

There is the further point not stressed by the authors: that international openness is a spur to productivity both because it allows for lower input costs and encourages innovation, and its omission is another reason why these estimates are conservative. Think of three parts of the argument. When you take tariffs off, the imports become cheaper for consumers. That is the first effect. The second effect, which the authors emphasized, is that it is not just that the price of Toyotas gets cheaper; the price of Fords gets cheaper as well, because Ford has to compete with Toyota in order to produce better cars. The third effect is that Ford has to up its game—which means that in addition to cutting its profit margins, it also has an incentive to innovate. We all do better when there's pressure to do better, and that leads to more productivity enhancement.

Finally, let me just reinforce the warnings that high inflation is a very substantial issue for the United States. Anyone who doubts that inflation is important should ponder that it brought Ronald Reagan, Richard Nixon, and Margaret Thatcher to power, to a very substantial extent.

The Fed's current forecast that we will be able to enjoy immaculate disinflation down to the 2 percent level with consistent unemployment at 3 ½ percent—I think it would be charitable to label it a hope rather than a dream and absurd to suggest that it is a likely prospect on the current policy path. That is because of the combination of excess demand and bad luck with supply shocks in recent years.

The key thing that I would advise people to watch is the data on wage inflation. Wage inflation in the most plausible estimates in the United States is now running above 6 percent and has accelerated significantly over the next quarter. If labor markets continue to generate 200,000 jobs a month, let alone 400,000 jobs a month, labor markets will only tighten from their current extremely tight level. That suggests the likelihood of wage acceleration. There are whole economic literatures directed at estimating the relationship between wage inflation and price inflation. I would boil it down to price inflation over time will be about 1 percent less than wage inflation.

If we are anywhere near 6 percent inflation rate, and at risk of that accelerating, we have a quite substantial problem. We need all the help that we can get in responding to it. The administration should focus on things that are real and important, like trade protection, not on things that are small and inconsequential, like dubious price gouging allegations. And that means heeding the result of this important Peterson Institute research on the benefits of trade barrier reduction.

Lawrence H. Summers is Charles W. Eliot University Professor and president emeritus of Harvard University. He has served as vice president of development economics and chief economist of the World Bank, undersecretary of the Treasury for International Affairs, director of the National Economic Council for the Obama administration from 2009 to 2011, and secretary of the Treasury of the United States from 1999 to 2001. He also serves as vice chair of the board of directors of the Peterson Institute for International Economics.

More From