The imperfect science of how much tariffs will hurt
Economists try to model tariffs’ impact but actually believe it could be much worse
No matter how fiery the rhetoric around trade becomes, one comforting factor is that most economists calculate the impact of expected tariffs on the U.S. economy will be surprisingly small. The problem is that these same economists concede that their models don’t capture the complexity of trade and supply chains, and personally believe the reality could be much worse.
The Trump administration placed tariffs of 25% on $34 billion in Chinese products on Friday, drawing a matching, retaliatory tariff from China.
These actions came on top of the tariffs the White House has already imposed on steel and aluminum imports from Canada, Mexico and the European Union at the end of May. And there could be much more to come.
The number-crunching models economists use to build their forecasts suggest the ultimate effect of the tariffs that have been put in place so far will have a minimal effect on the $20 trillion U.S. economy. Moody’s Analytics model, for example, shows that the tariffs will shave all of 0.03% off U.S. gross domestic product in the third quarter, rising to 0.1% next year. But Mark Zandi, Moody’s Analytics chief economist, isn’t as sanguine about the tariff effects as his model is. “If anything, we’re significantly underestimating the disruption this will have on the U.S. economy,” he says.
Take Harley-Davidson’s announcement late last month that it would shift production to Europe of motorcycles it has been exporting there to avoid retaliatory EU tariffs. A typical economic model might predict that European sales of Harley motorcycles would fall because of the tariff. It wouldn’t predict that Harley would wholesale move production of nearly 40,000 bikes it sells in the EU out of the U.S., which amounts to a bigger reduction in GDP.
The models economists use to analyze trade are a lot more complex than the ones in introductory economics textbooks, which predict bigger losses to consumers from tariffs than gains for domestic producers and the government. Economists’ more sophisticated models include second-order effects, such as what tariffs do to inflation and how the Federal Reserve reacts. But moves like Harley’s plans aren’t part of the models. And there are other potentially big factors that they don’t include. Stock prices, for example, aren’t in a lot of them, and if they are there aren’t as- sumptions about how negative psychology could affect prices.
That means the models could be missing a crucial element in how trade could affect the economy. The U.S. economy is especially sensitive to sharp stock sell-offs, which can lead consumers to curtail spending and companies to reduce investment and hiring. Bank of America Merrill Lynch economist Ethan Harris points out that some of the stock market’s biggest declines this year have come on days when trade tensions have escalated.
The problem, says Goldman Sachs economist Jan Hatzius (who excludes stocks from his model), is that predicting how stocks will react to tariffs (or just about anything else) is difficult. You can raise your expectations in your model about how much stocks might decline and create a negative economic reaction to tariffs. “But if you do that, in some sense you’re assuming the answer,” he says.
Other sentiment effects — how worries about tariffs might affect businesses’ plans or make workers worried about losing their jobs — are similarly tricky, notes Deutsche Bank economist Peter Hooper.
“Nobody’s model does a very good job of how uncertainty and hits to confidence affect behaviour,” he says.
The models also don’t address how in modern supply chains, where goods can be imported and exported multiple times through the manufacturing process, tariffs can be magnified as they come on top of one another. That could make the proposed up-to-25% tariffs on autos particularly onerous because car parts can cross borders multiple times. JPMorgan Chase economists point out that less than half of the content of cars sold in the U.S. is sourced domestically.
The U.S. economy is doing well, which will offset much of what isn’t captured in economists’ models. The danger is that if the U.S. continues to step up tariffs, and other countries continue to retaliate, the effects could be amplified.
Under Mr. Zandi’s model, if the administration proceeded with all the tariffs it has floated — the car tariff and the additional 10% tariff on $400 billion in Chinese goods — and China and other countries responded in kind, the hit to GDP next year would go to 0.5% from 0.1%. In an extreme case, where the U.S. put a 25% tariff on Chinese imports, and China responded in kind, the hit would rise to 1.3%. That is a big bite out of the 2.4% growth economists expect next year. Add all the things that aren’t in his model and the reality of any of those scenarios could be much more dire.