Inflation Undershoots Again in May —Tariff Effect Nowhere To Be Seen

June 12, 2019

By Jeff Ferry, CPA Chief Economist

This morning the Bureau of Labor Statistics (BLS) published the May consumer price index, and once again inflation has undershot Wall Street expectations. 

Headline inflation came in at 1.8% for May 2019 (compared to May 2018). The core rate of inflation (i.e. removing volatile food and energy prices) was slightly higher, at 2.0 percent, but still undershot what analysts were expecting. 

One team of bank economists reacted to this morning’s undershoot with a published note saying: “Feeding the frenzy—Softer US consumer price inflation data will only fuel the market clamor for interest rate cuts.” The note went on to speculate that Fed chairman Jerome Powell might signal a bias toward easing interest rates as soon as next week. 

Meanwhile, over in the world of trade, free traders continue to insist that tariffs are raising prices, even though there is almost no evidence of such price increases, and everywhere else in Washington, economists and politicians are grappling with disappointingly low inflation. 

So let’s look at the evidence. The steel and aluminum tariffs imposed in 2018 were allegedly going to raise the price of consumer products that use steel. Vehicles are big users of steel and aluminum. According to today’s Bureau of Labor Statistics news release, new car and truck prices actually fell 0.1 percent between April and May. Year-on-year, that price index is up 0.9 percent, meaning it is putting downward pressure on the nation’s overall core inflation rate of 2 percent. 

Recently, Goldman Sachs published a note claiming that inflation in tariffed industries was much higher than in non-tariffed industries. One industry they cited without providing actual data, was furniture. Well, the BLS has kindly provided the data, and it shows that furniture and bedding prices rose 2.4 percent in the year to May—slightly higher than core inflation but not enough to make a real difference. Indeed, Williams Sonoma, one of the leading suppliers of furniture in the US (via its Pottery Barn and West Elm divisions) recently reported an excellent quarter. Moreover, Williams Sonoma noted that it is moving its furniture production out of China, highlighting the fact that much of US industry isn’t sitting on its hands, it is working aggressively to reduce its dependence on Chinese production. 

Another comment on tariffs came recently from the head of the beermakers’ trade association, who has complained about the price of aluminum boosting the price of cans used for beer. Well, today’s BLS data shows that the price of beer consumed at home was up 1.9 percent year-on-year, slightly below the core rate of inflation. Furthermore, if you read the recent investor conference call transcripts of the CEOs of the leading US beer companies, you will find that there is zero discussion of the price of aluminum. In these calls, where the CEOs are legally required to focus on what they believe is affecting the course of revenue, profits, and their stock prices, the CEOs focused on their core issue, which is Americans’ declining consumption of mass-produced beer (in favor of wine, craft beer, and other fruity alcoholic concoctions), as well as the difficulty of ramping up supply for those brands that are doing well, and the increasing cost of transportation.

Finally, let’s look at washing machines. This is the one item where the anti-tariff crowd can find some statistical justification for their arguments. Washing machine prices did shoot up by some 10 to 15 percent in 2018, after Section 201 tariffs were applied to the industry. Yet in May 2019, laundry equipment prices are actually down 1.6 percent compared to a year ago, showing the prices are coming back to earth. In any case, washing machine prices are still well below 2012 levels. The artificially low price levels of 2015-2017 were depressed by dumped products from Asia. 

Looking at the big picture, goods inflation was actually negative in May, down 0.2 percent from May 2018. Services (less energy services) rose by 2.7 percent, pulling up the core rate of inflation. If we look within services, we can see the real drivers of inflation today. Rent, for example, was up 3.4 percent (all figures year-on-year unless otherwise specified). Health insurance rose a whopping 12.4 percent. Tuition, school fees, and childcare rose 3.5 percent. 

In other words, goods inflation continues to be weak as foreign competition puts downward pressure on goods prices, along with the wages of those US employees who work to provide those goods. Services inflation, although modest, is much higher than goods inflation because in general, services are insulated from low-price foreign competition.

Federal Reserve Bank of New York and other Works of Fiction

In light of this weak inflation report, what are we to make of claims by economists like those at the New York Fed that the costs of the tariffs are being completely passed onto US consumers? Their claims don’t match the data.

Here’s what the New York Fed said in its recent “study” of tariffs. First it argues that due to the 10 percent China tariffs, “U.S. domestic prices at the border have risen one-for-one with the tariffs levied in that year….the 2018 tariffs imposed an annual cost of $419 for the typical household. This cost comprises two components: the first, an added tax burden faced by consumers, and the second, a deadweight or efficiency loss.” The study then claims that when the tariff rose to 25 percent on China imports: “we estimate that the annualized deadweight loss increases from $132 to $620 per household, bringing the total annual cost of the new round of tariffs to the typical household to $831.”

That $831 figure has received undue publicity. It is highly misleading in a number of respects. Its most basic claim contradicts common sense. If the 10 percent tariff applied to many Chinese imports beginning in June 2018 was completely passed onto US households, as the NY Fed authors claim, why can we not see it in any goods at all containing China imports? Many products we consume every day contain some proportion of China imports, from $2 toys to $60,000 SUVs, yet none of these products show that sort of price increase. Indeed, many of these prices have fallen slightly. And if those prices are hitting the companies within the supply chain, why do so few of them complain of profits suffering from higher tariff costs? Some companies (Ford, GM, Caterpillar) have indeed cited higher costs due to tariffs, yet their profits in North America have not been affected. 

The answer is that the data that the NY Fed is using is flawed, and import prices have indeed fallen. 

Next, the NY Fed study argues that companies will switch production elsewhere and this will raise costs. They cite Vietnam as an example. It is indeed true that companies are moving production out of China and into Vietnam, even as I write this. However, studies we’ve seen, and anecdotal information from CPA members, tells us that costs in Vietnam are in general lower than those in China. Even with the costs of moving production to a new location, the ultimate costs are unlikely to be higher, so this element of the NY Fed’s costs is also unrealistic. 

The NY Fed study makes much of a concept beloved by trade economists called “deadweight loss.” This refers to the idea that if labor or capital is deployed in making a product that could be made more efficiently somewhere else, then there is a “loss” compared to what those resources could be producing elsewhere. This is an example of something endemic in trade economics: concepts that exist in theory but not in reality. We’ve already mentioned that moving production to Vietnam does not lead to a “deadweight loss” since costs are no higher than China. Another example is rising production in the US. The steel tariff has led to a surge in employment by steel mills all over the US. In Osceola, Arkansas, Big River Steel is paying an average of $75,000 a year to staff at the new mill. The NY Fed economists would argue this represents a loss to the US economy because Chinese steelworkers could do similar work for pay of around $10,000 a year (supported of course by billions of dollars in Chinese government subsidies, and the country’s lax pollution standards). 

But the NY Fed view only makes sense if the US workers could be employed in other industries and still earn, i.e. produce value, equivalent to $75,000 a year. In reality, the typical semi-skilled Arkansas worker outside of the steel industry is probably earning half that, or maybe less. In other words, the Arkansas economy is less productive and Arkansas GDP is lower, without the steel mill. The “deadweight loss” concept only works if all the resources from the lost industry can be redeployed at the same or higher rates of return as in the lost industry. That almost never happens. 

The final critique of the NY Fed and similar studies is their loaded use of English.  Consider this sentence: “Studies, including our own, have found that the tariffs that the United States imposed in 2018 have had complete passthrough into domestic prices of imports.”

The use of the past tense (“imposed,” “have had”) makes it appear that this supposed effect has actually happened. But what they really mean is that their model suggests that that is what happened. There exists an entire twitter feed called “Just say in DSGE models” which argues that economists should stop claiming things have happened when they mean their model shows that this is what might have happened if their model is correct. 

In conclusion, all economic model results should be carefully scrutinized to see if their projections match real-world observations.

Showing 2 reactions

Please check your e-mail for a link to activate your account.
  • Jeff Ferry
    Thanks Warren…I’m looking forward to the article where you explain this analysis in greater detail!
  • Warren Platts
    Excellent article Jeff. I would add that their main result—100% passthrough isn’t even statistically significant according to their own calculations. But if you back calculate the passthrough from their more reliable figures that were significant at p<0.01, then using their numbers with the largest sample size there is actually only a 56% passthrough entailing a terms of trade gain of $5.2 billion. Moreover, their estimate of the deadweight loss of $6.9 billion is exaggerated,.because it does not take into account the effects of moving production to even cheaper places like Vietnam. Once you do that, the deadweight loss is reduced to $4 billion. Thus, according to the standard net welfare calculation where national net welfare = the terms of trade gain – the deadweight loss, then there was a modest boost of $1.4 billion to national net welfare, even by Amiti et al.’s own numbers! And of course all these papers completely ignore the gains to national productivity that happens when people quit their jobs in the service sector to take better jobs in manufacturing. It is a sad state of affairs when you cannot take at face value 99% of what mainstream (English speaking) economists have to say on trade…