TAIWAN TO CHINA
Taiwanese CFOs position themselves on the mainland as politics make way for business logic.
By Abe De Ramos

Harvey Chang, CFO of Taiwan Semiconductor Manufacturing (TSMC), has the perfect resume for Taiwan Inc. First the head of his own securities house, then a major fund manager, he was hired five years ago by TSMC's power-broker CEO Morris Chang (no relation) to run finance at the foundry, by then the world's largest. He has helped build US$3.6 billion-a-year TSMC into one of ex-Japan Asia's truly global companies, a cash-rich powerhouse, and a draw for international capital.

It's hard to believe that a man like this would need more schooling, but his boss packed him off to the US for two months this year for an Eisenhower Fellowship program - a kind of "genius" course for international executives. The purpose of the Eisenhower sounds esoteric: to foster better international relations through business. But in CFO Chang's case the aim is practical. He's responsible for taking Taiwan's premier company into China, a country that Taiwan considers its greatest political and military threat.

Politics aside, TSMC needs China to survive. Taiwanese investment into China has become a crucial source of growth for China's economy and a staple for Taiwan business. Taiwanese manufacturers dot the mainland. Towns up the Pearl River Delta are nicknamed "little Taipeis". Despite the saber-rattling between the two countries, back-channels have always operated between the renegade and the giant. In the last ten years, various businesses from the island have pumped US$50 billion across the strait, often via third country entities to work around the legal, and highly political, hurdles.

Now a crucial door has opened. Until four months ago Taiwan forbade key technology investments into the mainland. With intense industry lobbying, the government reluctantly eased the ban, giving companies like TSMC direct access to the vast market. As a result, CFOs in Taiwan are about to benefit from the full potential of the mainland. A lower cost production base, once the key reason to move to the People's Republic, has now been supplanted by more urgent motives.

China, despite its poverty, is now one of the world's biggest markets for electronics. China has increasingly become the hub to supply goods to the rest of Asia, and Taiwan Inc. must retain access to this, one of the crucial areas it has targeted for sales growth. And most important, Taiwan's high-tech companies must be able to manufacture in China to help their customers, which have all built operations there, to cut costs. For example, TSMC's customers will be able to avoid a Chinese tax levy against imported goods if the company can make its wafers locally. "Many of the Taiwanese companies leveraged their China manufacturing base to provide a cost-effective approach," says Stan Shih, chairman of the US$3 billion-a-year Taiwan-based computer maker Acer. "Now China is very important for us to access the global market," he says.

Acer and its subsidiaries already make computer and related components in China, but easing of restrictions allows Shih to manufacture mobile phones and laptop computers. No sooner had the Taiwanese government lifted the ban, than Shih decided to build a plant in Jiangsu province to make notebooks under his contract manufacturer, Wistron. China's potential also gave Shih renewed hopes for the brand he spent decades to create, after flailing in the US (see box, "The Reinvention of Stan Shih").

Shih's statement resonates the belief of many Western multinationals that came to China to establish their claim of being global. But while tales of heartbreak, from car makers to beer brewers, abound, Taiwanese companies may actually prove to be more successful. For one, they are making leading-edge products that local Chinese companies are either just learning to make, or are not making at all. As such, domestic competition is muted, or at least level. Their timing is also right. China has just become a member of the World Trade Organization (WTO), and should gradually eliminate conditions that disadvantage foreign investors, from high tariffs on imported capital goods to forced technology transfer.

No doubt, technology companies in Taiwan are already global players, as most consumer brands in the world outsource many of their products to contract manufacturers there. But these outsourcers are inexorably crossing the strait. Almost all customers, potential customers and importantly, competitors, are increasing their presence in China, to serve not just the mainland, but the Asian region. This year alone, analysts expect US$4 billion in semiconductor facilities investments in China. Not surprisingly, the world's largest contract manufacturers - Solectron, Celestica and Flextronics - are also shifting some plants to China.

All are drawn by the lure of China's consumer electronics market. "With 150 million subscribers, China is now the largest market for mobile phones in the world, as it will soon be for a lot of other consumer electronics products, by the sheer size of its population," says Manoj Menon, technology director for US market strategy consulting firm Frost & Sullivan in Singapore.

All, too, have recognized that China has become Asia's major trading hub, almost supplanting the United States in volume of trade flow with Asian markets. Taiwan's major companies need Asian growth to satisfy their global ambitions, and the mainland offers the best - perhaps the only - option.

Tax Break

TSMC's only action so far was to open a sales office in Shanghai last year. Its chairman, Morris Chang, has been the most vocal and acerbic critic of the ban, knowing China would only be too eager to host his plants. Now, CFO Harvey Chang is evaluating locations, and should be seeking the approval of both Taiwan and China anytime in the second half of the year. "We are spending quite a bit of time looking at the pros and cons of the incentives that were already offered to us by the local governments there," he says.

TSMC is the perfect example of a Taiwanese company going to China for reasons other than a lower cost. Labor costs at TSMC amount to 5 percent of operating costs, so the savings the wafer giant might accrue in China are limited. Nevertheless, the pressure on TSMC and its competitor United Microelectronics (UMC), a close second to TSMC in terms of size, is still palpable, and that's because competition is beginning to build up.

The third-largest player, Singapore's Chartered Semiconductor, has struck a technology and capacity alliance with Semiconductor Manufacturing International (SMIC), China's first advanced foundry. On their own, local foundries are increasingly capable of spending big, as access to domestic and foreign capital is easy given the potential size of the market. Newcomer Grace Semiconductor, for example, is investing US$9 billion over the next ten years to build five fabs in its 360,000 sq-meter site.

These foundries are already sharing the same customers as TSMC and UMC, and their orders are increasing. It may be ridiculous to consider them as threats to the Taiwanese giants now, but given that they were not yet even in business three years ago, their capacity for growth should not be underestimated. Growth is, in fact, almost certain. Integrated design manufacturers (IDMs) - or companies that design, make and market chips on their own - are increasingly outsourcing the making part, and SMIC has had some recent wins from Japanese accounts, including Fujitsu.

Chang knows the outsourcing trend will only continue. Currently, foundries - or companies that make chips for design houses and IDMs - make up just 15 percent of total chips in production, while the rest are made by the IDMs themselves. But US-based Applied Materials, which supplies the equipment that makes semiconductor chips, thinks the foundries' share will go up to 25 percent by 2004. Other estimates point to 50 percent by 2010. "Eventually, IDMs will probably account for 40 to 50 percent of our client portfolio, from below 30 percent now," says Chang. "We see that continuing, because not a lot of semiconductor companies can afford the [next generation] 300mm [wafer size] fab, and they will gradually move to the fab-lite, or even fab-less direction," he says.

Because China will be the second-largest market for semiconductors in the world by 2010 or sooner, many IDMs and design houses are, as expected, selling to China. TSMC can only benefit from establishing its own fabs there. The most urgent reason: value added tax (VAT). China slaps a 17 percent VAT on imported semiconductors, as opposed to 3 percent for locally manufactured ones. For now, the local fabs can only supply 25 percent of total demand, so vendors have little choice but to import from TSMC and UMC. But the two companies have been warned.

"Our US customers, who have either joint ventures or wholly owned subsidiaries in China, have indicated to us that sooner or later, we have to be there, because it costs them a lot in taxes to import our goods," says Chang. "It's probably okay for a while that we just ship to them, but later on, we have to be there because there's also going to be a wide spectrum of users of our products," he says.

So Far Away

Those users point to the second reason TSMC has to be in China. Making chips for design houses and IDMs requires close and immediate contact with their engineers, something that could not work for TSMC if the engineers are mainland Chinese. The political strife with Beijing has kept Taipei wary of illegal migration from China to Taiwan, and the visa application process is prohibitive. It takes mainland tourists up to two months to get a visa. Business travellers have it worse: if they are engineers, it could last up to six months. In the fast-evolving chip industry, that is unacceptable.

"For foundries, the most important success factor is time to market," says YC Lin, chairman of the supervisory board of the Taiwan Semiconductor Industry Association (TSIA), and head of operations at ProMOS, a Taiwanese maker of DRAM chips. "It takes face-to-face discussions with the engineers to examine the quality of the product, and if the engineers can't get visas, that will ruin their cycle," he says. TSIA has been actively lobbying Taipei to ease these restrictions, but the government remains unwilling, knowing it is a deterrent to the migration of talent and technology. "As soon as [Taiwan] lifts the transportation ban, they can reduce the motivation of companies to move to mainland China," says Lin.

With a plant in China, Chang will not have to worry about this issue. However, other restrictions mean that TSMC will not get the full benefits of its own technology. The lifting of the investment ban only applies to 200mm wafers, a lesser technology than the new 300mm wafers that TSMC is ramping up. The larger wafers are supposed to save buyers as much as 30 percent (as there will be more dies per wafer), but for importers in China, this will be undermined by the 17 percent VAT. TSMC and UMC can relax, though, since Chinese companies are just learning to cope with 200mm wafers. "China is pretty much in the development stage, but it's too early to tell if they will never be able to catch up," says William Dong, analyst at UBS Warburg in Taipei.

For now, Chang is glad to be on track with TSMC's ten-year, US$20 billion investments in 300mm wafers, despite dismal earnings in 2001, the worst year in semiconductor history. The Taiwanese government requires that 300mm wafers be in volume production before TSMC can ship its 200mm wafers to China. Currently, 300mm wafers account for less than 5 percent of TSMC's total production, but he expects mass production as soon as next year. By that time, Chang will have freed enough 200mm equipment to be initially capable of making up to 30,000 wafers a month in China, about two-thirds of SMIC's projected capacity.

Of foreign competitors coming to China such as UMC and Chartered, Chang says: "We see the semiconductor business in China gradually mature, and certainly other people are taking the same view. They're trying to have a first-mover advantage, but I think it's a matter of who will be first to go to the market, rather than racing for the location."

Dong says it could take up to six months for a product to be transferred from design to production to stable yield, and given its veteran's experience in 200mm wafers, TSMC can ramp up quickly. "China has demand for some of the larger market share products. Actually, most of their demand is still for more mature technologies, like smart cards," says Chang. "We'll gradually grow from one manufacturing site; it's possible to have two, three or four plants," he says.

So Chang has been shuttling between Hsinchu and Tainan in Taiwan, checking on each fab and reading yield reports, to see which equipment he will ship to China. "We're talking a lot, not some," he says. Why is Chang so gung-ho about China? Simple. The CFO faces almost no financial risk. Capital investments will be small because used equipment will be shipped. And if Chang is financing US$20 billion for 300mm plants in Taiwan out of internal cashflow, then the millions that TSMC will spend to build plants in China are trivial.

With only a 0.20 percent debt ratio, TSMC is underleveraged. "We don't borrow much, so most of our investments are self-financed," says Chang. "Our business is very cyclical. If you're highly leveraged, and you go to a slowdown period, you make your debtors very, very nervous," he says.

Going Solo

Unlike some existing foundries, TSMC will go to China not through a joint venture but as a wholly owned subsidiary. Intellectual property rights (IPR) is one reason. Although China under WTO is supposed to strictly enforce IPR, the risk remains. Recently, TSMC filed a suit against a former employee for using its intellectual property for a fab in China. "China has IPR laws, but the enforcement is as good as in Taiwan," says an analyst with a European investment bank in Hong Kong. "That is one reason why somebody wouldn't want to give his key IP. These are soft issues, but they do matter," he says.

Chang says, however, that the more important reason for the structure is management. "The semiconductor industry is very dynamic, and we'll be able to make decisions much faster as a wholly owned subsidiary," he says. This is crucial especially when expansion plans, which require billions of dollars, are to be made. "If you have other partners, it will take you time to communicate, and they don't necessarily always understand," says Chang. "We're not talking about a small amount of money, so I think the pace would be slowed down," he says.

TSMC's and UMC's forays onto the mainland are not necessarily unwelcome for foundries already operating in China. Frank Lai, CFO of Central Semiconductor Manufacturing (CSMC), sees their future investments as more beneficial than competitive. CSMC is a joint venture between a Chinese electronics company and a US-incorporated venture capital fund backed by a Taiwanese chief executive. It makes 150mm wafers that will not directly compete with the new Taiwan heavyweights. One of Lai's concerns is a lack of support infrastructure for testing and packaging chips. As such, some wafers that CSMC makes in China are sent to the Philippines for packaging, and then to Hong Kong for testing, and then re-imported to China. This extends the cycle time. "When the Taiwanese start to come in, you will see an influx of upstream and downstream support to make the whole industry viable," he says.

With its eyes firmly on China, TSMC is not just going to help spur the domestic chip industry, but it will also be a keen player in the sector's eventual consolidation. After all, his Eisenhower Fellowship is on mergers and acquisitions.

At the end of the day, China's relevance to TSMC will not outshine the benefits of Taiwan. For one, Chang laments the lack of engineering talent, a point on which Lin of TSIA agrees. "There are plenty of high-quality personnel for entry level, but for the mid-level managers, it's very challenging," he says. For example, NEC Hua Hong, a Sino-foreign joint venture with NEC of Japan, had to move 100 to 200 Japanese engineers to the station in Shanghai. "In terms of people cost, I don't think there are any savings, because to send someone overseas, you've got all kinds of benefits and packages."

True, China produces 100,000 engineers a year, but the core research and development work comes from engineers who studied in the West and came back to work for Taiwanese companies, and China has few of them. "There's a lot more willingness for the Taiwanese who come from the US to return to Taiwan, than there is for the Chinese to go and work in China," says an investment analyst. "I'm not saying they won't be able to get engineers, but they'll need to compensate competitively," he says.

Two in One

The importance of going to China, on the other hand, is double for Eric Yu, CFO of BenQ, which makes computer peripherals and mobile phones. The reason: BenQ has both a contract manufacturing business and its own brand. As such, Yu has to be equally sharp in competing against other contract manufacturers - mainly fellow Taiwanese and the North American giants - and against local and foreign brands at the same time. On the contract manufacturing side of the business, Yu's formula for making money is by saving it; for the BenQ brand it's by spending on advertising and promotions.

While BenQ's immediate business growth is sure to come from the contract side, more dollars are being funnelled to its brand campaign in China. Yu's belief: if you can make it in China, you can make it anywhere in Asia.

It's a bold, ambitious strategy for a name that's hardly readable (it's pronounced Ben-Q). But looking at its brightly colored, Benetton-inspired ads that border on the risquZ¹ in conservative China, its determination is undeniable. The good thing is, BenQ may be right. "It's a good strategy, because by being in the biggest market you can achieve economies of scale, which will help the company in many ways," says Menon of Frost & Sullivan. One is being able to bring down prices, and be very price competitive in the market. "Second is, if you are selling more, and are able to reinvest a lot of this back into research and development, you'll be able to sustain and continue the business growth," he says.

With four years of experience in manufacturing under its former name, Acer Multimedia and Communication, and another four developing the BenQ brand, Yu is not exactly struggling in China. BenQ is currently the hottest vendor of optical devices such as CD-ROM, DVD-ROM and CD/RW. It also leads the market for keyboards and LCD monitors, and trails respectably at number three or four in CRT monitors, projectors and scanners. Now, there is no looking back. "Taiwan is a very small market, so our focus is greater China," says Yu. "Then we will build our reputation in Asia Pacific, then later Europe and the US," he says.

Now, BenQ is marketing its mobile phones, which, thanks to the government's lifting of investment restrictions, it can now make in China. BenQ has been making them for Motorola in Taiwan. As early as the last quarter of 2001, it has started transforming certain production lines in Suzhou, an industrial park just outside Shanghai, to accommodate mobile phones. From zero, BenQ now estimates at least 30 percent of all its mobile phones will be made in Suzhou. China's demand for mobile phones means this isn't a small number. Yu expects to double its total production from 7 million last year, to 14 million this year.

As a contract manufacturer that relies heavily on labor, BenQ is the perfect beneficiary of everything cheap that China has to offer. The salary of one R&D staff in Taiwan will pay six in China. The lifting of investment bans - which also removed the annual cap of US$50 million a year - means Yu is less eager to pursue further investments in Penang, where 10 percent of its revenues last year were made. Before, BenQ financial controller Alex Liu said he kept a policy to keep revenues from China at less than 30 percent overall, "to manage political risk. After WTO, and now that China's regulations are [more favorable] to Western companies, we'll increase that portion."

That is, dramatically. Of its projected revenues of US$3 billion this year, 50 percent will come from China, from just 30 percent last year. Thanks to its contract manufacturing business, which because of mobile phone demand will account for 70 percent of revenues this year, his production cost is very low. To support its volume growth in China, Yu plans to invest US$20 million a year in the next three years.

BenQ's cost savings do not end with setting up shop in China; it's also beefing up its R&D to explore further cost savings. One handset, for example, used to have 250 components inside. BenQ is bringing it down to 200 by integrating certain components to reduce the parts count, which will then make it easier and cheaper to produce.

Unlike TSMC, Yu is confident of the R&D quality in China, because the applications it needs for mobile phone operations are less demanding than silicon chips. "Taiwan is also short of human resources, and aside from being high-cost, most of the people there are not willing to work overtime," says Liu. "This industry fluctuates a lot, sometimes there are huge orders coming, so we are forced to do overtime," he says. Adds Yu: "China can make us faster."

No one will probably disagree with that statement, and not just colleagues from Taiwan, but other countries as well. As politics give way to business logic, the CFO is always the most immediate beneficiary.

Abe De Ramos is a senior writer at CFO Asia based in Hong Kong.