Europe's Crackdown on Chinese EV Exports

Mike Smitka
Prof Emeritus of Economics, W&L Univ
Judge, Automotive News PACE Awards for supplier innovation
Steering Committee, GERPISA global auto industry research network
Contributor, SeekingAlpha

Europe is poised to launch an antidumping countervailing duty investigation of Chinese EV exports to Europe. Launching is a political decision. Once the "go" is given, however, it turns into an administrative process that runs according to set rules. What matters is whether the lawyers involved find subsidies in China; they will. [You can find many details here on the European Commission Directorate-General for Trade website, including the link to a spreadsheet of recent EU cases.] It won't matter whether the EU provides similar incentives. Once launched, a guilty verdict is certain, and that could well lead to punitive tariffs that will bring an end to European EV imports from China. The process, however, will take several months, so might not affect European imports until June 2024.

The initial announcement focused on the threat from cheap EV imports from China, naming several of China's EV startups. But that's not how the process works. The EU plays by the rules, and those rules focus on an industry or product category, and in general not an individual firm. So the reality is that the real focus, intended or not, is Tesla. That's because they're the dominant EV exporter – all of Europe's Model 3s come from Shanghai, and many Model Ys. Stopping those exports from Tesla's Shanghai plant will not only lessen the pressure on the German car companies and their high-priced EVs. It will also force Tesla to turn up the volumes of production of the Model Y in their plant in Berlin.

Other EU firms will be caught in the cross-hairs. The Dacia Spring EV comes from China; Dacia is Renault's lower-price brand. Some BMWs and Volvos come from China. Their volumes are still modest – the Spring sold just over 5,600 cars in August – but their plans are expansive. Tesla's volumes, however, aren't trivial. Over the past 12 months Tesla sold 103,000 made-in-China Model 3. I don't have detailed data for made-in-China Model Ys, but I conservatively estimate half of the 200,000 Tesla exported the past 12 months made their way to Europe. At 1 million units a year, that output represents 20% of their Shanghai plant's capacity. In other words, without exports to Europe the Tesla plant in Shanghai can't surpass the 80% utilization level that is the auto industry rule of thumb for being profitable. Making cars, after all, comes with high fixed costs, and the irony is that the lower the wage level, the more plant-level fixed costs matter. Tesla now sits in the normal position of a car company with aging product: cut prices to maintain sales and capacity utilization or keep prices high and see sales plummet. (There is after all what the media calls a price war going on in China's EV market, touched off by none other than Tesla.) Either way, margins collapse. Fortunately Tesla's starting point is one of good margins, so that alone won't push the company into the red. But it does undermine the case for investing in Tesla's stock.

I add further detail in my Sept 26 article on SeekingAlpha. First, I provide underlying data. One key observation is that the era of rapid growth for China's EV market is over. Market expansion will no longer allow EV makers to see sales expand despite keeping prices high. Tesla has launched a refresh of the Model 3, produced for now only in Shanghai. The Chinese auto press however pans it as not worth the most prominent new feature, a higher price. And things will look much worse if the slowdown in the Chinese economy expands from real estate and shadow finance merchants in China's lower-tier cities to the wealthier metropolitan areas such as Shanghai where most EVs are sold.

Addendum: I continue to look into the process, and previous EU countervailing duty examples. The calculated subsidy will be the sum of national level subsidies such as reduced income taxes, provincial subsidies and local subsidies. Those individual items of 1% here and 0.5% there can add up to a significant total, the 19 line items for Chinese truck & bus tires add up to 21.97% (min) and 75.91% (max), so the bottom line in that case is substantive. The DG-Trade does not substantiate every claim, or finds some below the threshold to do detailed calculations. And the geographic variation of policies in China mean that the final CV duties will likely vary from firm to firm, as the industry is spread across a wide swath of China.
There are many issues. How will they treat the free license plates for EVs in 8 of China's largest cities? – I have no idea. If data aren't provided, what will they use as a proxy? One example of subsidized land for a factory in China uses the price of a similar sized site in Taiwan as the comparator. That may or may not be a higher price, but there's some incentive for the industry in China to cooperate lest the Trade Directorate choose arbitrary comparators that are unfavorable. However, recent Chinese legislation that classifies much data as "sensitive" may impede cooperation, lest company officials end up in jail for sharing classified information with a foreign government.
Out of the 432 line items in the database, 274 are for purported subsidies to exporters in China; many others are for India. As with antidumping cases, the steel industry generates multiple examples, but there are also ones for fiber optics, for commercial vehicle tires and on and on. br />Finally, one commentor on SeekingAlpha who has a Europe background believes the political decision has already been made. So now the devil will lie in the details.

Original article Sept 26, the final paragraph and other modifications addedd Sept 27.

China's EV Market: 2023Q2 update plus an analysis of Guangzhou Auto and its Aion EV brand

Mike SmitkaProf Emeritus of Economics

In July the Guangzhou Auto (GAC)'s Aion brand was #2 in China's EV market, behind BYD but ahead of Li and Tesla. However, the EV market is no longer expanding rapidly, and macroeconomic stormclouds suggest the normal fall sales surge will be muted. Still, does investing in GAC make sense?

In an article published today (Sep 2, 2023) on the finance website SeekingAlpha, I argue that GAC will not continue its rapid growth, while the ongoing restructuring of its recently shuttered joint venture with Mitsubishi Motors should improve profitability. Two other joint ventures, with Toyota and Honda, also appear to be profitable, but again don't provide a growth story – they're only just adding EV assembly lines, so are also capacity constrained.

Management states that Aion is currently profitable, and that they anticipate an IPO. If that comes to fruition, it would make GAC more attractive, and might be interesting on its own. But that would await the ability to show a full year of profits, so I believe won't take place until next spring.

Now SeekingAlpha articles are normally behind a paywall, but you can try going here China 2023Q2 NEV Update: Focus on Guangzhou Automobile and Aion. My analysis draws on a database I've created over the past 3-odd years that has monthly sales data by model since Jan 2020, plus drivetrain, price bracket, brand, firm and segment. I read articles from the Chinese-language auto websites on a daily basis, and of course have overall background from 3 decades of research on the Japanese and North American auto markets.

Automotive Productivity: Plant Architecture

Mike Smitka
GERPISA Steering Committee
Judge, Automotive News PACE Awards
Prof Emeritus of Economics, W&L

This is the first little bit of an article on factory architecture, why it changed and its implications for the industry

I've both enjoyed and leared a lot from the substack articles at Contruction Physics. I've started thinking about how I would approach the same general set of topics for automotive. Starting on the construction end of things, early factories were fairly compact, because the use of steam power required locating things as close as possible to the power plant. The first factory I worked in was huge in length, to the point that bicycles were still used, but it was also built in 6 floors. If you look closely at the Diego Rivera Detroit Industry Murals, based on Rivera's first-hand observations in 1932 of Ford's Rouge complex in Dearborn, Michigan, you still see belts in use. Electric drive motors were pervasive by that time – Henry Ford worked at Detroit Edison prior to launching his first of three successive automotive startups in 1903. The transition, however, took decades.

Ford's operations when he started building the Model T were on the top 2 floors of a small building in what is today the Ford Piquette Avenue Plant in Detroit, a half mile north of the Rivera Murals at the DIA. It was an open hall with a series of work stands. At that time Ford relied on craft techniques, carting components made by outside suppliers up a freight elevator to be assembled by experienced fitters into a vehicle. A railroad spur ran up to the back of the building, essential for bringing in bulky and heavy engine castings and frames.
Now a museum, it makes a wonderful venue for receptions – we held the closing "gala" of the June 2022 GERPISA conference there, surrounded by Model T versions and "upfits" into pickup trucks, snowcats and so on.

Ford's first purpose-built factory, a few miles away in Highland Park, didn't come until 1910. It was a multi-story operation, but unlike Piquette was full of machinery, and from 1912-3, a moving assembly line.

China's Population Decline: Not "News"

Mike Smitka
Judge, Automotive News PACE supplier innovation awards
Steering Committee, GERPISA auto industry research network

Prof Emeritus of Economics, W&L

That China's population would start declining in the 1990s was a "done deal" a quarter century ago. By that point the One Child Family policy had extended the existing low fertility regime of China's cities into the much more populous countryside. The resulting smaller families meant that by 1998 the much smaller cohort of potential mothers under the One Child policy were themselves having few children. Meanwhile, urbanization meant that the number of families for whom a single child made sense was rapidly expanding, while a countryside where child labor was central to the well-being of families was fading. Between changes in target fertility for young mothers, and the cumulative impact of below-replacement fertility, future population decline was locked in place.

...that's China's population decline makes headlines is a surprise...

A bit more on the details. Maintaining a stable population requires that women average 2.1 surviving children, to ensure that there's at least one daughter to replace each mother. Since 1980, estimates of the "Total Fertility Rate" have remained well below 2.0 – even with the policy's many exemptions, families with 3 or more children remained a distinct exception. Hence a quarter century ago it was already clear that China's population would at some point begin to decline. Not only was the number of 15-19 year old women 6% lower than in 1980, the number of 0-4 year girls was 15% lower. By the mid-1990s the age of marriage had also risen, and teenaged mothers are now highly unusual. So by 1998, the mothers of the late 2010s had already been born, and were far fewer in number. Even if the timing of onset and rate of decline couldn't be pinned down without knowing how rapidly lifespans would increae (ditto subsequent fertility changes), the broad outline was there to be seen by anyone who understood basic demographics.

China's Population Structure, 1998
from PopulationPyramid.net

Of course we now know those details. Longevity did increase, and helped by very low child mortality, a girl born today can expect to live 80 years. Older Chinese are in no hurry to die! On the opposite end, however, fertility hasn't just remained low, it's fallen. As a result, the number of 0-4 year girls in 2022 is an estimated 40% below the number of young 20-24 year old women in 1991. Even if these girls go on to average 2.1 children – so far all evidence points to them not doing so – the number of mothers will fall through 2043. As the population pyramid shows, with today's girls the smallest cohort in decades, total population wouldn't start to stabilize for another 60 years, when that cohort of women starts to die off. Because China's population is so large, an option available to the US or the EU – accepting large numbers of immigrants – won't work. There simply won't be enough young people elsewhere in the world to fill the gap.

China's Population Structure, 2023
from PopulationPyramid.net

I've simplified the story. Not all women have children at age 20, and mortality rates vary with age and can predictably be expected to change as today's younger Chinese grow up with cleaner air and better drinking water, and are far less inclined to smoke. Of course they're also better fed and sedentary, which works in the opposite direction. Fertility rates will be different, even if the experience of Japan, Korea, and Southern Europe suggest they won't rise. Chinese exhibit a preference for sons, and by the late 1990s families engaged in selective abortions, so that the ratio of boys to girls rose 10 percentage points; the TFR thus needs to be 2.3 to guarantee that enough young girls will be born to replace their mothers. Aging likewise is uneven, tempered by the timing when larger cohorts reach their 70s. So demographers combine data on age-specific fertility and mortality to generate projections. Pandemics aside, those are pretty good over a 2-decade timespan. Thereafter the high-fertility, low-mortality case diverges from the low-fertility, high-mortality one. Still, the fact that every mother in 2043 has already been born provides a solid baseline.

That the likely population decline of 2022 makes headlines is thus a bit surprising. That the media coverage fails to focus on the consequent decline of the working age population, which is of the essence to future economic growth, is even more surprising. I know though from 3 decades in the classroom that the demographic momentum that comes from a population's age structure and birth/death rates is not intuitive. Even I, more numbers oriented that most humans, was once surprised by how much we know two decades in advance. It's also key to understanding retirement issues, and that too is not intuitive to a 20-something. As our societies age, though, it will come to dominate all of the middle- and high-income societies.

Addendum

That having a second child is viewed as challenging comes out in a January 24, 2023 New York Times article, "They Poured Their Savings Into Homes That Were Never Built" by Isabelle Qian and Agnes Chang. The article traces 4 individuals who took out mortgages to purchase apartments on which construction has ceased. Three indicate demographic implications: one a broken engagement and end of plans to have a child, and two who now "... can’t imagine trying to buy another home or having a second child," and a second who proclaimed that, due to the lost downpayment on a storefront property and continued mortgage payments, they are “... afraid to have another child. The income and expenses barely break even.” Nothing a government run by old men is going to do will enable such couples to have the three or more children necesary to reverse population decline.