Carney: How the Fed Messed Up Its Tariff Study

The New York Fed recently released an analysis of the costs of tariffs on the U.S. that rests on an error so fundamental that it undermines the entire project.

The Fed economists argued that the Trump administration’s tariffs must be costly to U.S. households because they assume they impose “deadweight losses.”  Here’s how they explain those losses:

Some firms may also reorganize their supply chains in order to purchase their products from other, cheaper sources. For example, the 10 percent tariffs on Chinese imports might cause some firms to switch their sourcing of products from a Chinese firm offering goods for $100 a unit to a less efficient Vietnamese firm offering the product for $109. In this case, the cost to the importer has risen by nine dollars, but there is no offsetting tariff revenue being paid to the government. This tariff-induced shift in supply chains is therefore called a deadweight or efficiency loss.

In other words, the analysis rests on the assumption that it is more expensive to manufacture outside of China than within China, despite the well-known fact that China’s labor costs have risen above those in many other nations. Why do they assume this? Basically, it is a kind of efficient market hypothesis. If manufacturing is taking place somewhere now that must be because that is the most efficient place for it.

Whatever sense that seems to make, however, collapses under the weight of a moment’s thought. Just as the Fed shows that U.S. tariffs could theoretically move manufacturing into less efficient countries, China’s pre-existing tariffs have almost certainly driven manufacturing out of more efficient locations.

This can be illustrated with a simplified example. Imagine a world with three countries, two large developed nations and one small developing nation.

Let’s say it costs $100 to make a widget in Oceana, $110 to make it in Great Plains, and $90 in country New Delta. In a free-market, the widget will be manufactured in New Delta.  For the sake of simplicity, assume that all of the widgets get exported to Oceana and Great Plains and that manufacturers face frictions and management costs that mean they must choose one country and only one country to produce goods.

If you are in a free market environment and you notice that widgets are made in a New Delta, it’s safe to assume that it is the low-cost manufacturer.

But that assumption doesn’t work if you introduce tariffs. If Oceana imposes a 15 percent tariff on imported widgets, it no longer makes sense to manufacture them in New Delta for export to Oceana and Great Plains. Those products will now cost $103.50 in tariffed Oceana. So production shifts to Oceana to avoid the tariff.

This means that in the presence of tariffs, it is no longer safe to assume that widget production in one country signals that country is the low-cost manufacturer. Oceana’s tariffs have imposed a $10 deadweight cost on the rest of the world.

So what happens when Great Plains puts a 25 percent tariff on imports from Oceana? Production shifts back to New Delta to avoid the higher tariff. In other words, the retaliatory tariff imposed by Great Plains has shifted production to the lower-cost manufacturer. The price of the widgets made in New Delta is actually lower. Instead of raising prices, tariffs have lowered prices.

Tariffs have eliminated a deadweight cost instead of imposing them.

This is not just a theoretical exercise. Prices of imports fell in May when the U.S. hiked tariffs on many Chinese good up to 25 percent. Many companies have said they are shifting production outside of China. The Fed’s deadweight costs are not evident in the data.

China, for years, imposed an average tariff of more than 15 percent on imported consumer goods, while the U.S. tariffs were approximately zero. For goods sold both in the U.S. and China, the tariff disparity drove production into China that might otherwise have gone to other countries.

The Fed is right to say that tariffs and trade barriers distort economies and can inflict deadweight costs and efficiency losses. The mistake was to begin the analysis assuming that the U.S. tariffs were the first trade barriers introduced into global trade rather than just the latest.