How do I love thee, Huahong?

We talk a lot about China’s industrial subsidies, particularly in the tech sector. (I know I do - here’s my overview piece on what I call China’s ‘Disruptive Industrial Strategy’). They are at the core of (whisper it softly, cadres) Manufacturing 2025. And right now, US Trade Representative Robert Lighthizer is on a righteous crusade to curb them. (Though, of course, he’ll fail, in no small part because his boss has decided the re-elect is more important). But, where exactly do the funds come from, and where do they go? And how should the USTR aim to curb them? I thought it would be worth laying out, in some detail, just one example of how public money flows into China’s semiconductor space. And I thought I’d chose a firm which is not even well-known nor cutting-edge, and still gets showered in taxpayer love, to reveal the scale and shape of what’s going on. Introducing; Hua Hong Semiconductor Ltd. Or Hua Hong, for short.

Back in the good old days, you could clearly see SOE subsidies in a line in the Ministry of Finance’s budget. Now, these were not usually funds designed to foster tech-upgrading; more often they were a life-support money drip. Today, something similar goes on, but more quietly. SASAC, the SOE ministry, recycles most of the dividends it receives from SOEs right back to SOEs, keeping the stragglers alive with the monopoly profits of the strong. Some Chinese economists who should (and probably do) know better say that since these MoF SOE subsidies have been eliminated there are no subsidies anymore. It would be hilarious if it wasn’t so demonstrably untrue.

Tech sector subsidies are way more complicated, endemic, and heterogeneous than a single budget line.

The subsidies available break down into many different pots. There are central and local government subsidies; the former tend to be strategic industry-focused, while the later are the crude incentives cities use to attract capex. There are loans from government-run banks as well as “equity injections” from government funds (which are in turn funded by loans from government-run banks). There are income tax breaks, subsidies tied to R&D investment as well as sales revenues, and, for the particularly desperate city, there are “We’ll make sure you don’t make an operating loss” promises as well. And these are just the supports I’ve found for Hua Hong.

Sorry, I should have introduced you properly.

Hua Hong is a Shanghai-based, Hong Kong-listed, lagging-edge semiconductor firm. ‘Lagging’ because its current fabs (factories) produce chips using the 55nm and 40nm node processes, much larger (and thus less sexy) than the 7nm that leading-edge foundries like TSMC are now using. (In short, the smaller the node process, the more circuit you can print onto a smaller surface, the more energy efficient and (ultimately) cheaper the chip is.) So Hua Hong makes relatively simple chips. It makes the CMOS image sensor chips in your smartphone cameras. It makes Bluetooth chips, power management chips, microcontrollers (MCUs, the tiny CPUs which run your fridge and TV), and the chips in your ATM and ID cards. Some two-thirds of Hua Hong’s sales end up in consumer products; electrical appliances, cars, those cards. It was running at close to full capacity in 2017-18 in the chip boom. (Which ended about a year ago - since then demand for most chips has faded with the auto, smartphone and even ‘hyperscalar’ cloud investment cycles slowing.) About half of its output is sold in China; the rest is exported. Thus, what happens at Hua Hong does not stay at Hua Hong - it affects the global market.

Hua Hong is the listco and controls the PRC vehicle, HHGrace, which operates all the onshore fabs. Hua Hong, in turn, is controlled by the HuaHong Group, which in turn, is 47.1% owned by China Electronics Corp (CEC, 中国电子信息产业集团公司), a SASAC-overseen central SOE, and 47.1% by Shanghai United Investment Fund (SAIL, 上海联合投资有限公司). SAIL was set up by Jiang Mianheng, elder son of Jiang Zemin, out of the Science & Tech office of the Shanghai government, in 1994 here. It is thought by some to be still influenced by the Jiang family (I could not possibly comment). SAIL also has a 15% direct stake in Hua Hong. One executive I met recently claimed that that strong political backing is one reason the firm is allowed to focus on making money, such as that is, rather than, like its Shanghai neighbor SMIC, being forced to invest in much riskier, leading-edge nodes. But more on that later.

Hua Hong currently has the world’s second largest 8-inch (200mm) wafer capacity. Wafers are what you etch your semiconductor circuits onto – and the leading players work on bigger wafers (12 inch/300mm). More on that later too.

Thankfully, Hua Hong, as a listed entity publishes its accounts, so there is some transparency about its financials. Business as usual involves a number of subsidies which run through its income sheet. For the budding forensic accountants out there, here are the details…

There are two over-arching forms of support, mostly from the Shanghai government, I believe. First, funds which are not related to Hua Hong’s daily operations, which just appear on the income statement as cash. These totaled USD 24mn during 2013-17. Second, and bigger, there are funds which are linked to the firm’s activities – of which there are three types:

Here, the accountants book the government’s grants as liabilities on the balance sheet, and these are then released/reduced over time as the company choses to recognize the support. I guess in theory this generates some incentives for the firm to sell stuff and invest in R&D, but still, I wonder if they’re very strict. Perhaps more importantly, Hua Hong can decide when it releases the funds, which allows it, to some extent, to manage its annual earnings. Injections of these grants totaled some USD 75mn over the same five year period. So, that’s about USD 100mn in 2013-17, USD 20mn a year – during a period when the firm was making about USD 100-200mn a year in net income. That’s Shanghai tax-payer money.

But that ain’t nothing compared to what happened a couple of years ago when Hua Hong decided to build a new fab.

In August 2017, Hua Hong announced it was going to build a new fab in Wuxi, a city 120km away from Shanghai. It’s their first 12-inch plant (a big upgrade from 8-inch), but still focused on 65-90nm processes. Its nearly built, equipment is being installed, and they should ramp up in production later this year. Fabs cost a lot of money, and so are risky investments at the best of times. Phase #1 of this one will cost USD 2.5bn. But when the site is fully built out, the price tag could be nearer USD 10bn. This project one looked risky even before semi demand started to weaken last year; the world is crying out for more leading-edge capacity, but another lagging-edge fab is a tougher sell. Maybe we’ve hit peak smartphones; credit cards seem suddenly rather 20th-century (at least outside of the US), and the much-trumpeted 5G roll-out (and any IoT revolution which accompanies it) is in the future.

But Hua Hong does not seem that bothered, as the Wuxi venture is structured to minimize any negative impact to its finances. Make sure you’re sitting down for this, as the Wuxi fab subsidies are ridic.

First, Hua Hong itself got an USD 400mn equity injection from China’s National IC Fund (which is funded mostly by China Development Bank loans, I believe) in November 2018. That gave it the funds it needed, as well as a big backer in Beijing. The fund, known in the sector as “the Big Fund” (大基金), is a weird beast; partly commercially-oriented (they say they want to make money), partly national-strategic (they don’t), and is run by a chap who has very limited experience in chips. (The Big Fund deserves a deep dive in itself…maybe another day.)

Hua Hong then established a joint venture with an investment vehicle controlled by the Wuxi City government. Hua Hong (Wuxi) was established in February 2018. Hua Hong took 51% (and injected USD 918mn, about a half of which came from those Big Fund monies), and thus control, while China’s Big IC Fund directly injected USD 522mn, taking a 29% stake. And Wuxi Xihong Lianxin Investment injected USD 360mn (20%) in cash.

The total investment is USD 2.5bn, USD 1.8bn in equity, and the rest (USD 700mn) in bank loans, mostly again I’m told, from China Development Bank. These loans were at LIBOR +180bps, I’m told, which sounds very attractive, and were five+ years in term.

So, in short CDB was all over the deal – injecting funds into Hua Hong itself, as well as into the Wuxi subsidiary, and then lending to the new venture to boot. I do sometimes hear myself introducing CDB as China’s World Bank - but, err, I should probably stop doing that. Maybe, “SoftBank’s Vision Fund with public savings” would be better. Anytime CDB gets into trouble it gets a nice capital injection from the FX reserves.

But there’s more! Wuxi’s officials found some nice land in its hi-tech zone, and provided it at a low price. But, of course, this is hardly unusual; industrial land is usually given away by cities needing to hit their fixed investment goals and generate future tax revenues.

But that’s not the kicker. The Wuxi government guaranteed to subsidize Hua Hong Wuxi’s operations “for the first few years”. The terms of this agreement have not been made public (which is a bit frustrating, I imagine, for shareholders in Hua Hong), but it seems Wuxi has promised to match any operating losses at the new fab for some amount of time. So, if the new venture loses USD 20mn a year for three years, then the Wuxi tax-payers are on the hook for USD 60mn. And those taxpayers’ funds flow straight through the accounts into shareholders’ pockets. Heads, Hua Hong wins; tails, Wuxi’s taxpayers lose! (And what if the fab is a huge failure - will Wuxi stump up USD 120mn?)

Last year, Wuxi’s tax base was CNY 86bn (USD 12bn). So maybe those subsidies won’t be noticed. While Hua Hong (Wuxi) will not create that many jobs, it is possible that it generates a supply chain in the vicinity, which would be job+tax positive. Its worth bearing in mind, though, that the city inevitably already have debts as well as PPP projects for which space on the future budget will have to be found. I doubt anyone has done a budget out beyond 2019.

And, almost forgot, as a high-tech firm, Hua Hong already benefits from an average income tax rate of 16%. As a result, it’s been paying some USD 20-40mn in taxes a year recently.

Now, of course, Hua Hong is not the only tech firm taking advantage of local governments desperate for a spanking new fab.

BOE (京东方), the LCD/OLED panel maker, is notorious for nurturing competition between local governments to compete for its new plants, and reportedly requires them to stump up at least 60% of the cost of a fab. So, for instance, in Hefei, the city promised CNY 6bn (USD 1bn) and ended up forking up CNY 12bn (USD 2bn) for a CNY 17.5bn (USD 2.5bn) sixth-generation LED plant, according to [this piece] ( Bloomberg recently ran a very nice piece on BOE (here). The firm is still expanding, most recently in Chengdu, where no doubt the city is throwing money at it too. And as a result of financial support from numerous cities and banks, plus aggressive learning of the technology, BOE can now make foldable OLED screens for Huawei’s new phone. That’s quite something - and should petrify Samsung, which is hanging on by its fingernails to the #1 slot in OLED production.

You’re probably already super-familiar with how the incentives work at the local level. A party secretary and his mayor have three things at their disposal: land, considerable flex on tax and endless charm and wit. (I might have got one of those wrong.) Localities revolted when MoF minister Lou Jiwei attempted to eliminate such local tax flex a few years back. Officials can also often influence local banks, pushing them to lend to new ventures. (One important route is the ability to decide which banks get to look after the city’s tax and land sales funds.) Ironically, though, it kind of makes sense for banks to lend to government-backed fabs and such, given they know the borrower has strong government support.

So, Wuxi’s officials swapped some land and tax revenues for a big new shiny investment. Fabs are expensive; so they represent maximum kaohe zhibiao points for the amount of effort invested to attract the investment. Future tax breaks are the successors’ problem (Yep, a similar game to local government debt! As I explained in this piece focused on nearby Zhenjiang. It’s not clear, though, yet, how well Wuxi’s leaders will do out of the deal. Li Xiaomin, the Party Secretary since 2015, is still there, and looks set to retire into the provincial congress. The vice mayor who ran the Hua Hong project, Huang Qin, is now mayor, though. So he’s done alright.

Which brings me to three final thoughts.

First, its clear that the Hua Hong Wuxi fab is only being built because of state funding; lots of it, from many different sources. It is not a commercial venture. Some of its is tax-payers money; but most of it comes from a state-controlled bank. Its USD 2.5bn, which is a lot of money - but Hua Hong is an after-thought compared to the leading edge efforts that folk like YMTC, SMIC and Innotron are up to. Now, maybe you think there is nothing wrong with that - China should have the right to build such capacity, especially, as many folk will now say, if the US is willing to cut off tech supplies. I’m sure if you asked one hundred Chinese folk on the streets of Beijing what they thought about that, all of them would agree.

But the policy is purely mercantilist - it is all about ensuring production takes place in China - and is exported. With endless state support, BOE will eventually destroy Samsung, Hua Hong will kill Vanguard, its closest US competitor. This is not free trade - however much Beijing’s leaders claim to be upholding the WTO - and other countries have the right to complain about it, and react. I view tech export controls from the US as a reaction to China’s policies - though its clear we’ve now entered a vicious cycle whereby the more the US and others control technology, the more Beijing will double down on its disruptive industrial strategy. We’re not heading to a good place.

Other argue that China’s disruptive industrial strategy will end in failure; wasted savings and sub-par technology. I’ve explained what I think about those arguments here.

And there’s an interesting minority view that the world should be happy about China lowering the global price of MCUs, as it did with solar panels etch. Hmmm. Well, if we’ve learnt anything about trade in the past few years, its that consumers are not the only folk we need to care about; workers are important too. And there is another thing we now need to care about, unfortunately…

Second, there are the national security ramifications of China attempting to absorb the entirety of the semi-conductor supply chain. Hua Hong will not be producing anything particularly exciting at its Wuxi fab; it’ll be more like the bread and butter of chip-making. But this stuff is important too when you think about all the millions of simple chips that go into IT systems, industrial systems, telco systems as well as weapon systems, these days. A policy of not trusting one’s neighbors for chip supplies inevitably raises questions about how friendly you’re intending to be in the future.

Third, Mr. Lighthizer and his fight against industrial subsidies. Of course, unless he’d have got his way and tariffs were continually raised until Beijing really felt some pain, he was not going to get very far. His boss, after playing a blinder in 2018, is now showing his true pathetic colors and is ready for a superficial deal on soy and gas. Pork and hot air would be an apt tombstone for this president.

But even with a half-decent president, this stuff is complicated. Maybe one might have found a way to bring a WTO case against Shanghai or Hefei offering subsidies. But going after government-backed “industrial funds” making investments is way more complex. That’s not covered under WTO rules. Maybe USTR could have insisted funds such as the Big Fund should be shuttered in the negotiations? China would not have agreed, I’m sure. Or maybe they could have insisted that their funding only come from private investor sources. Good luck on getting the detail on that tied down.