Discover more from The Asianometry Newsletter
TSMC’s Decision to Build in China
Author’s note: You can watch the video before reading the script:
The debate to enter the mainland is worthy of a Harvard Business Case Study. Economically, I think the decision was a success. TSMC Nanjing is a dominant player on the mainland and in some estimates is the third largest foundry by itself. The company took some pretty thorough measures to prevent its data from leaving the company. Starting the Nanjing and Shanghai Fabs also helped checked growth of SMIC and other fab upstarts on the mainland. I made a video about SMIC that briefly touched on this.
But I think that was always what people thought would happen. TSMC is too big and dominant not to have made it a short-term financial success . The question is whether or not it was the right decision to have made in the long run.
I cannot speak for TSMC, but I suspect that there were two reasonings: The first was that in the long run, their technologies were going to fall anyway to the competition. There is no such thing as a company that can stay five years ahead of the competition without continuing to push the limits of technology. It is definitely a more stressful way of running a business - but it does work.
The second is that the company wants to be a neutral zone when it comes to geopolitics. They just want to make and sell chips and they will do it for anyone they legally can. With recent developments in current events, such a balance might be a bit more difficult to strike. But for now, the company can serve mainland customers with its highest end processes.
If you like the content that the Asianometry YouTube puts out and want to support the channel, feel free to check out the Patreon.
In 1999, the People's Republic of China entered the World Trade Organization and changed the world economy forever.
The exploding Chinese market attracted many Taiwanese investors and policy makers saw a potential threat to their industries.
Foremost in these concerns surrounded semiconductor manufacturing, the crown jewel of the Taiwanese high technology industry.
This video is a bit different today since I am going to run it like a university case study. I will pose to you a question and give some information. Think about what you would have done in that situation.
This decision came at a turbulent time in Taiwanese politics. In 2000 Chen Shui-bian and the DPP ended 50 years of KMT rule and ascended to the presidency of the Republic of China.
Just a few years ago, the DPP had been an illegal opposition party. Now their man was in the Presidency.
However, like US President Barack Obama, he presided over a divided government. The KMT continued to dominate the legislature and as a result the two clashed over almost everything.
President Chen campaigned on the back of a pro-environment and pro-welfare policy. He demonstrated his commitment to this policy with the expansion of the Taiwanese budget and in stopping the construction of Taiwan's fourth nuclear plant.
Chen also continued the path of his KMT predecessor Lee Teng-hui, enacting broad cultural and educational changes with the goal of remaking the Taiwanese society as something of its own rather than as just another Chinese province.
Much of the Taiwanese business classes felt uncomfortable with these changes, contributing to the continued capital flight across the Strait. Taiwan's economy soon began to struggle and the unemployment rate rose to 5%.
(Yeah I know, 5% is not high by American standards. But it is for Taiwan.)
The Taiwanese dilemma
Up until the early 2000s, Taiwanese foundries were limited to just $50 million of investments in the mainland. This is due to a 1991 rule put in place by Lee Teng-Hui called "Don't rush. Be patient".
The rule has a nice smile but sharp teeth. Investors who did not report their cross-strait investments could be fined or imprisoned. UMC president Robert Tsao would be subsequently prosecuted under this law.
The question then was whether or not the Taiwanese government should allow its semiconductor manufacturers to establish and build foundries on the mainland.
In making this decision, they had to consider many different factors.
In 2000, two new semiconductor foundries opened in China - sending shock waves across Taiwan and kicking off this debate.
The first was Hongli or Grace Semiconductor Manufacturing Company. It had two big name co-founders:
Mr Wang Wen-yang, the son of the founder of the Formosa Plastics and Formosa Petrochemical industrial giant.
And Jiang Mian-heng, the son of Communist Party General Secretary and Paramount Leader Jiang Zemin.
These two names alone was enough to shake the foundations of the Taiwanese semiconductor industry.
The second new Chinese startup was Zhongxin or Semiconductor Manufacturing International Corporation (SMIC). It was founded by a former TSMC vice president, representing the first of many major cross-Strait defections.
These upstart semiconductor foundries used near unlimited bank accounts to fund incredible amounts of employee poaching. This remains a problem to this day, with mainland firms offering compensation packages of 4-5x what Taiwanese firms can afford to pay.
It is not unknown for these employees to bring over entire teams to China ... as well as their proprietary technology.
Right then UMC and TSMC did not consider these startup foundries to be serious competitors in the market. They will need years and years before developing into such. But there is no better way to kill off a competitor than in the cradle.
Author’s Note: TSMC utilized this method to snuff out a few other competitors including Chartered Semiconductor. Check out the video about Singapore’s semiconductor industry
The massive market
TSMC and UMC wanted to have first mover's advantage in China and take major share of the market - serving local integrated chip designers there. The Chinese market is brand new and offered infinite possibilities.
Frank Huang, then CEO of PowerChip Semiconductor, said:
The China market is growing. Taiwanese companies can’t miss out on it—we have to be there. There’s no reason that the foundries should not be there. The ranking officers in the government all know that
TSMC CEO, Chairman and founder Morris Chang argued this point most forcefully - saying that the Chinese market would grow 20% a year for the next two decades.
China was willing to allow these Taiwanese foundries into that fantastic market, but there was a catch.
As Chang said:
TSMC’s policy is to keep its headquarters, R&D center, and manufacturing business in Taiwan, and to market around the world, except some special places such as China. China is a conservative market. We need to manufacture the wafers there to enter the market.
In other words, they had to give China something in order to get something. Their technology, for the Chinese market.
Protecting their technology
Technology transfer is a sensitive subject for any nation. Taiwan would know this lesson better than most others. After all, it was technology transfer in the late 1970s that helped Taiwan enter the market in the first place.
In 1976, Taiwan put out the call to American, European and Japanese companies for help in establishing a domestic semiconductor market. RCA won the bid out of 7 companies and trained a team of 37 engineers for over a year.
So essentially the United States allowed some of its most advanced technologies to flow to Taiwan for $10 million. It was a different time politically and economically, though certainly the friendly relations between Taiwan and the United States helped.
Now Taiwan faced the same potential situation, except they were on the other side.
Taiwanese foundries argued that they would take steps to protect their most advanced technologies - transferring over older stuff. This would be their 8-inch wafer technology as opposed to their cutting-edge 12-inch wafers.
In any case, US technology export restrictions prevented TSMC and UMC from transferring their latest fabrication tech - which at the time was 0.25 micron or smaller. Yes, even back in 2003 the US were restricting technology transfers to China. It's not a new thing.
The last factor I want to bring up has to do with jobs.
I don't want to sound like a Marxist here - but management and labor have different goals and views on certain subjects. Management might be super excited about establishing foundries on the mainland but their employees have different ideas.
Taiwanese employees vociferously voiced their opposition to the idea. They saw it as management's attempt to pay them less and to invest less in Taiwan. Investing less in Taiwan would mean less jobs for Taiwanese in Taiwan, forcing them to go abroad to find work.
The Taiwanese Engineers Union estimated that the transfer of 8-inch wafer technology would cost Taiwan 17,000 jobs over the next 2 years. The union Vice President led a small protest of a 1,000 engineers, saying:
We have lost 6-inch [production facilities] already. We are going to lose 8-inch and then 12-inch. How much more can we lose?
This is indeed the real impact of losing what management calls "a less-advanced technology". It means the loss of people unable to have opportunities to learn how to walk before they can run.
You would just instead have a bunch of highly paid executives and specialty engineers at the top and a bunch of poorly paid people at the bottom - the hollowing out of the middle class, kinda like how it is in many American companies.
This was a tough decision. I listed a majority of the main factors people were considering, but I certainly missed a few others. Because this is history, you can easily Google it and figure out how it decided. And in a few seconds I will tell you.
But before then, I would like to pause and ask you, dear Viewer: How would you have decided if you were the Taiwanese government?
In the end, the Taiwanese government decided to allow the establishment of the 8-inch fabs on the mainland. This is with certain conditions - the foundry first had to invest a substantial amount of money to upgrade to being capable of making 12-inch wafers.
The decision was extremely divisive, and it made for some strange bedfellows. DPP President Chen and his ministers supported it.
But Lee Teng-hui opposed it.
As well as the head of the Mainland Affairs Council - a then-unknown Tsai Ing-wen. She said that lifting the ban on wafer investment was like releasing two tigers:
"Once these two tigers eat up the rabbits in China, they will come back and eat the Taiwanese rabbits."
To be honest I am not quite certain what that means but it sounds cool.
Public opinion drifted from being 50-50 in support of lifting the ban to 27% in support (with another third saying that they did not know enough on the issue). The opening up of the Taiwanese semiconductor manufacturing industry would mark the start of friendlier times in Cross-Strait relations - a time that would be further supported during the subsequent administration of KMT President Ma Ying-jeou.
For the foundries, the decision was a financial boon. In 2007, TSMC would build their first fab on the mainland in the Chinese city of Shanghai. Their second fab would be opened in 2017 in Nanjing.
Thus far, these fabs have been very successful. The Nanjing fab alone generated half a billion dollars in revenue in 2019 - which basically makes it by itself the third largest foundry on the mainland.
But by then, relations between Taiwan and the mainland chilled. President Ma had left the scene as one of the least popular presidents on record. Tsai cruised to a 2016 victory.
Were the naysayers right? Did it help or hurt the Taiwanese economy? Economically, it is hard to find support for this in the data. Some sectors gained and others lost - but that is how it is for any economy in response to any situation. The data has been extremely messy.
In the end, it might be best to say that the move helped Taiwan economically and financially, but at a political cost. Many companies got richer. Improved relations with the Mainland is better for business, at least in the eyes of the Taiwanese business class, and also ironically gives them the assurance to continue investing in Taiwan.
For many Econ majors studying demand supply charts in college but not much else - that should be the end of that, right? Higher GDP = success!
But such an analysis disregards societal and political costs. And both were extremely high. Taiwanese businesses continued to invest in China, and the issue got caught up in others relating to the struggling economy. Wages stagnated. House prices continued to soar (and still do to this day). Young people kept moving to America or the mainland. The Taiwanese people saw continued struggle - culminating in the momentous Sunflower Student Movement in 2014.
Know anyone who likes the topics that the Asianometry newsletter goes into? Feel free to share it with a friend using the button below