“I am a Tariff Man,” declared President-elect Donald Trump. But wait! Won’t tariffs cause inflation? Yes, say the Wall Street Journal and other mainstream commentators. No, says Scott Bessent, the hedge fund manager that Trump has picked to be Treasury Secretary. At most, he thinks tariff inflation will be a bump in the road.

Who is right? The inflation experienced after Covid-19 offers some clues. The supply-chain disruptions that set it off were only transient, but the resulting inflation had a distressingly long tail. Three lessons learned from that painful episode – which may have cost the Democrats both Congress and the White House – suggest that any inflation driven by tariffs on the scale Trump has promised will be much more than a bump. 

Lesson 1: Inflation has both a demand side and a supply side

Start by dissecting Bessent’s Panglossian view, as revealed in a recent radio interview. “Tariffs can’t be inflationary,” explained Bessent, “because if the price of one thing goes up, unless you give people more money, then they have less money to spend on the other thing, so there is no inflation. … Inflation comes through either increasing the money supply or increasing the government spending, and that’s what happened under Biden.”

There is a smidgen of truth in this, but just a smidgen. Yes, policymakers can control inflation by using monetary and fiscal policy to moderate excess demand – that is, spending power chasing scarce goods – but these tools works best if excess demand is the origin of the inflation in the first place. The post-Covid inflation was different. The latest studies show that demand played only a small role in the upward surge of prices that began in the winter of 2021. Supply-chain disruptions played a much larger role. Tariffs, too, would mostly cause supply-side inflation.

When faced with supply-driven inflation, whether caused by factory closings and shipping bottlenecks (à la Covid-19) or tariffs, it is not enough just to hold the line on monetary and fiscal policy. To fully control inflation, the Fed would have to substantially crank up interest rates, preferably while Congress cut spending and/or raised taxes. In that case, we might get the Bessent result in which decreases in some prices offset increases in others. But such a strategy would come at the cost of falling real output and rising unemployment – even a major recession. Not what Bessent had in mind. 

Lesson 2: Some prices are flexible while others are sticky

Not all prices move smoothly in response to changing market conditions. As I explained in an earlier essay written while the post-Covid inflation was still raging, some prices are “sticky.” That is, instead of reacting immediately to changes in demand, they rise or fall only sluggishly. Price stickiness would again play a significant role in any coming tariff-driven inflation, so the story is worth a brief recap.

According to data from the Atlanta Fed, the most flexible prices are those of goods like oil and wheat that are traded on commodity exchanges where they can change by the minute. Meanwhile, the stickiest are the prices of services – think city bus fares or college tuition – that come under review perhaps once a year, if that often.

The Atlanta Fed publishes separate indexes of sticky and flexible prices, using a cutoff that puts about half of all prices in each group. The flexible index is dominated by goods like food and energy, which happen to be the items that are most directly affected by tariffs. The sticky index, in contrast, consists of services such as medical care, public transportation and rents that are rarely traded internationally.

The impact of tariffs is further complicated by the fact that flexible-price goods are inputs in the production of sticky-price services. For example, a tariff on imported vehicles would quickly boost the price of new cars and trucks, which are inputs for providers of transportation services. However, it takes time for a higher price for new vehicles to feed through to increases in bus or Uber fares. Across-the-board tariffs, such as those proposed by Trump, would give an immediate upward push to a wide range of flexible goods prices. Most of any offsetting price reductions of the kind Bessent envisions would have to come from sticky services.

Lesson 3: Prices are more flexible upward than downward

Once we get away from stuff that trades on commodity exchanges, prices tend to be more flexible (or less sticky) in the upward direction than downward. Think of a car dealer who is quick to raise prices when inventories are low, but who cuts them only reluctantly when the lot is overflowing. Wages, which are one determinant of prices in every sector, are notoriously more flexible upward. It is rare for workers to turn down offers of higher pay, but they strongly resist wage cuts – sometimes openly with strikes, sometimes by staying on the job but working sullenly (and less productively).

This asymmetry of price flexibility is a major reason the Fed sets its target inflation rate at 2 percent rather than zero. Doing so allows it to accommodate minor changes in relative prices without the need for any absolute price decreases simply by allowing some prices to rise by more than 2 percent while others rise less or not at all. However, a 2-percent inflation target would not be loose enough to accommodate a major supply shock like a 20 percent across-the-board tariff. That would require absolute cuts in the (usually sticky) prices of non-traded goods and services – something that would be both disruptive and time-consuming.

This Is Not Mere Theory

In case you think all this talk about supply shocks and asymmetrically sticky prices is just econo-babble, think again: we’ve just been through it. As this chart shows, the behavior of flexible and sticky prices differed dramatically during the inflation induced by post-Covid supply-chain shocks.

Before Covid-19 struck (early 2020), overall inflation was consistent with the Fed’s 2 percent target, with flexible and sticky prices increasing on average at about the same pace. Flexible prices fell a bit during 2020 under the impact of skyrocketing unemployment and fears of more, while sticky prices barely budged. Then, as soon as economies around the world began opening again, supply-chain bottlenecks developed – remember those headlines about congested ports and shortages of shipping containers? Those sent flexible prices soaring on many of the same goods (everything from Chinese washing machines to polo shirts from Bangladesh) that would most likely be hit by the new tariffs.

As inflation got underway, sticky prices rose, too, but much more slowly. The resulting gap between flexible and sticky prices, which peaked in June of 2022, was unsustainable. Since flexible-price goods, like gasoline for transportation and meat for institutional cafeterias, are costly inputs for sticky-price services, the service sector struggled to break even and hold onto its workforce. The gap has fully closed only recently, some three years after the original post-Covid supply disruptions. As the chart shows, most of the gap was closed through a gradual increase of sticky prices. Flexible prices did fall a bit, but they did not come down nearly as fast as they had gone up. Above-target inflation did not end until the gap in relative prices closed.

Read It and Weep

Major shocks to relative prices, whether caused by a pandemic or a turn toward protectionism in trade policy, thus take years to work their way through markets. Could the Fed and the Treasury, working together, extinguish tariff-driven inflation completely while relative prices adjust? In theory, yes. But doing so would require a big dose of fiscal and monetary austerity in which unemployment would rise, real GDP would fall and a prolonged recession would ensue. The adjustment would require absolute decreases in the prices of non-traded services. And unless the tariffs were only temporary, the adjustment would take even longer than the adjustment to the short-lived post-Covid supply-chain disruptions.

Will we actually run this experiment in real life? I hope not. If we do run it, though, I hope that everything I have said here proves to be wrong.