By Jeff Ferry, CPA Chief Economist
The New Year dawns on an America where many manufacturing sectors are in a state of aggressive growth.
In December, America’s largest steelmaker, Nucor, announced it will add a coil paint line to its Arkansas mill. That line will have the capacity to paint 250,000 tons of steel a year, about 10 percent of the mill’s total output. It will mean the addition of about 50 jobs, on top of the hundreds of jobs Nucor is currently adding with greenfield mills now under construction in Missouri and Florida.
In November, kitchenware and furniture maker Williams-Sonoma announced a 6.4 percent increase in third quarter revenue and profit up 7.4 percent, driven by the huge success of its new retail brand West Elm, up 14 percent in the quarter, and in spite of 25 percent tariffs on most of its product lines sourced from China. As part of its commitment to get out of China, Williams-Sonoma is opening manufacturing in the US and adding employees in manufacturing and logistics.
And in October, the largest US manufacturer of solar panels, First Solar, announced the launch of production at a new factory in Lake Township, Ohio, its second US factory. An investment worth over $1 billion, the new plant has created 500 new jobs. At the launch, First Solar’s CEO praised the “relentless energy of our people” for enabling the company to meet its target launch date as announced 18 months ago.
The evidence from the American heartland is clear. The Trump administration tariffs are working. These companies are examples of a broader trend. As we’ve reported before, US steel companies are investing some $13 billion in new steelmaking or mills across the US. Around a dozen solar panel and solar cell manufacturers are following First Solar’s example and launching US-based production.
Economists See the Cloud Not the Silver Lining
Frustrated by the flow of good economics news, the economics profession looks for other means to wage their religious war against tariffs. Last month, two Federal Reserve Board economists, Aaron Flaaen and Justin Pierce, released a report, Disentangling the Effects of the 2018-2019 Tariffs on a Globally Connected US Manufacturing Sector, which claimed to show that the Trump tariffs have led to a “relative” loss in manufacturing employment in affected sectors, and other small but negative effects from rising input costs and retaliatory tariffs by China and in the world steel industry. This is the third recent academic paper that used regression analysis to attempt to demonstrate that tariffs are raising prices with little or no impact on the real economy.
For those who want better insight into academic economics, I would recommend another paper, Teacher Effects On Student Achievement And Height: A Cautionary Tale, by economists Marianne Bitler of UC Davis and three colleagues. In many school systems, it has become popular to reward teachers by teacher performance, on the basis that teacher performance has been shown to be positively correlated with student test results. This entertaining paper demonstrated that with New York City school system data, teacher performance is also positively correlated with student height, a measure that teachers “cannot plausibly affect.” In plain English, having a good teacher does not make you taller, yet statistical analysis shows a clear link—one almost as strong as the link between teacher performance and math and English scores.
What is going on? The answer, to quote an economist friend of mine is: “If you torture the data long enough, it will confess to anything.” Today’s computers are so powerful, and there are so many available additional variables that can be added to estimation equations, that a determined economist can always find a statistically significant relationship if he works hard at it.
It’s the Unobserved Variables
As Bitler comments, the likely explanation for implausible correlations is that “unobserved variables” have influenced the result. What does that mean?
Here’s what we know about the US economy in the early days of 2020: there was a manufacturing boom in 2017 and 2018. This boom lifted manufacturing employment by 190,000 jobs in 2017 to 12.545 million and another 264,000 jobs in 2018 to 12.809 million. That 2018 increase was the largest in 20 years. The Federal Reserve’s manufacturing production figures support this, showing manufacturing production up 2.5% and 2.2% in 2017 and 2018 respectively. Further, reports from leading US companies for those years support these positive results, with revenue and profits rising strongly at many manufacturers.
In 2019, the manufacturing sector experienced a pause. Fed data show that as of November, manufacturing output was down 0.8% compared with November 2018. However jobs are not down. Manufacturing jobs in November 2019 came in at 12.865 million, 0.4% above the year-end 2018 figure.
The 2019 slowdown could have many causes. One likely cause is the series of interest rate increases the Fed imposed in 2019 (and has since reversed). Another likely cause is the slowdown in foreign economies, including China and Germany. A third likely cause is simply that the surge of growth in the two previous years was so strong that the economy was due for a pause. Growth is rarely linear.
The fact remains that the US economy continues to grow faster than our major competitors. China is suffering more from the “trade war” than we are. In October, our imports from China were down 23 percent at $40 billion compared with the same month in 2018.
Why did the Flaaen-Pierce study find a “relative reduction in employment” in tariffed sectors? A couple of explanations spring to mind. The first is that the sectors of China imports that the Trump administration chose to levy with Section 301 tariffs have tended to be those associated with heavy manufacturing and so more exposed to the 2019 slowdown than the overall economy. That is an example of an “unobserved variable” cited by Bitler. If the regression analysis does not have access to the relationship of a sector to end-demand, it will do its best to explain it with something else, such as tariff effects.
Another possible explanation is raised by economist University of Wisconsin economist Noah Williams. In a paper published last month “Faulty Data Driving the Manufacturing `Recession’ in Pennsylvania and Wisconsin,” Williams argues that the employment data from the Current Employment Survey (CES) significantly understates manufacturing employment as compared to the bigger, broader and more reliable quarterly survey also carried out by the Bureau of Labor Statistics. Using the quarterly data for June 2019, Williams found 19,700 missing jobs in those two states alone. The Flaaen-Pierce study used the CES data, because it is more recent and being monthly, more congenial for regression analysis.
And that spotlights yet another feature of the recurring “academic” studies that economists have rushed out to condemn Trump administration trade policies. Two things that the academic economics community agrees on are (1) free trade is wonderful and (2) Donald Trump is awful. For many academic economists, the temptation to rush out a study purporting to show that free trade is great and Trump’s trade policies are terrible is too great to resist. For some, this could mean mention in the New York Times and perhaps even a tenured position at a university (in which case one is set for life).
They simply cannot be bothered to look at the actual data and reports from the hundreds of companies that have reported positive results in the past three years. This is another attraction of regression analysis for academics. It means an economist need not know anything about the real economy or the real industries upon which he opines. And by tacit agreement within the academic community, regression-based studies are only ever critiqued by other regression-based studies, so all is for the best in the best of all tenured worlds.
This is not to say that the tariffs are perfect. There are cases where downstream industries are suffering. But in a world where China has created completely artificial and highly subsidized prices in order to take over world industries, the answer is not to repeal the tariffs. Instead, the answer is to extend protection to those downstream industries so they too can enjoy growth by serving American customers and hiring American workers.
And while there are industries where prices have risen following the tariffs, there are many product categories, like steel and solar panels, where prices have actually fallen after the imposition of tariffs. And even in industries where prices have risen, such as furniture, they have risen by far less than the tariff rate of 25 percent. And while some companies (Wayfair) whose business model is predicated on selling cheap stuff from China have suffered from those tariffs, the more versatile and nimble companies, like Williams-Sonoma, have done very well despite tariffs.
And this is a final point that academic economists, with their simplistic mechanical models of trade and tariffs miss: that US companies are run by thinking, resourceful human beings. Tariffs stimulate US production and to the extent they do raise prices, importing companies have found a way to offer US consumers good or better value for money while absorbing the tariffs and riding the growth trends, and meanwhile the tariffs are delivering prosperity to more Americans via the growing number of manufacturing jobs.
Williams-Sonoma aims to eliminate all dependence on China by the end of 2021, as Williams-Sonoma Chief Financial Officer Julie Whalen explained to an investor in October. “We plan to get out of about half of our exposure in China by the end of next year … I think you can see that we’ve done an incredible job of offsetting this tariff exposure and this quarter, in particular, is a great example of that. When you look at sequentially improving gross margins with almost double the tariff impact and we’ve done that with the strong health of our business and our cost negotiations. We’ve been working on this for a long time.”