On the outskirts of Shanghai, in an otherwise undistinguished cluster of warehouses, a mechanized army is assembling. The units are massive, heavily armored, and capable of incredible feats of endurance and strength. They can bend steel, hoist buses and cut through metal with their hands. Fortunately for us, they come in peace. The current batch of industrial robots waiting on the loading docks at Kuka Robotics’s Shanghai distribution center is not here to enslave the human race, but to build it some cars.
For Chinese manufacturers, this army represents a solution to several vexing problems among its working-age humans, including shrinking supply, rising wages and higher turnover. It also offers companies a chance to move up the value chain by increasing productivity and precision. But for the warehouse laborers who gathered around to gaze at a gargantuan mechanical arm mounted on the back of a heavy truck, robots also represent an unwelcome source of competition. Robots can lift and carry things around warehouses too, and these days they can even drive. In July, a Chinese robot built by the National University of Defense Technology chauffeured a research team from Changsha to Wuhan down a high-speed expressway – through traffic.
Worse, industrial robots are cheap, and given that manufacturing wages are projected to keep rising by 20-30% for the next two years, they are going to get relatively cheaper. Tom Zou, chief operating officer of Kuka Robotics, gave an example: He said that one arc-welding robot costs about RMB300,000, while a senior arc welding worker costs around RMB100,000 per year, including insurance. But the robot will work 24 hours a day, making it two to three times as productive as the human. “That makes the robot’s cost almost equal to hiring three workers for one year, except that normally you can use a robot for 50 years,” he said.
Just when everything was looking up for Chinese labor – better health insurance, better security, and higher wages – suddenly the bosses seem to be preparing to swap them out for machines that work harder for longer, cost less, and die indifferently in the traces without a whimper.
For the industry, of course, this is all good news. Burgeoning demand led by Asian economies has kick-started a rapid
recovery in the US$15.4 billion global industrial robotics industry, which experienced a precipitous 47% decline in orders in 2009. According to statistics from the International Federation of Robotics, orders for robots used in assembly and manufacturing increased by 27% in 2010, and are projected to continue to do so through 2013 by an average of 10% per year. Around half of them will work making automobiles, and most of the rest will work in building home electronics, but there are also niche applications, like packaging food (robots tend not to carry disease) and carrying things around warehouses.
Ge Wenjie, who covers the machinery sector for Nomura Securities, said there’s plenty of room for growth in China, especially in the automotive sector. “Look at the number of welding robots installed for every 1,000 automobiles. In Japan in 2008, there were roughly eight robots per thousand. In Germany, 6.6; in South Korea and America about 5.6. In China, there are only two.” He pointed out that as Chinese automakers move up the value chain to cater to an upward shift in domestic demand, their investment in robotics will increase disproportionately; high-end automobile manufacturing requires a higher ratio of robots.
In anticipation, robot makers are quickly expanding their operations in China. In March, Germany’s ABB announced the completion of an industrial robot manufacturing plant in Chengdu capable of producing 150 units per year. Zou of Kuka said his company is planning to follow suit in 2012, building a new manufacturing facility near Shanghai.
The investment is part of Kuka’s worldwide push to expand desperately needed production capacity from 15,000 units per year to 25,000. But the new China plants are not for outsourcing production to serve other markets – Zou said China offers few cost advantages in robot manufacturing anyway. Rather, Kuka wants to shave precious weeks off delivery times for Chinese manufacturers, who are clamoring to automate but cranky about industry-wide delays and order backlogs.
The industrial robot sector appears to have been caught flat-footed by the pace of the recovery. “When the recession struck, everyone downed tools and said, ‘We’re not going to sell anything.’ So they didn’t order enough parts,” said Andy Kirkwood, managing director of UK-based Global Robots, which refurbishes and resells used industrial robots. He said that the supply bottleneck helped make last year his company’s best ever; he sold 450 second-hand robots into factories, mostly in the US and Europe but also a few in China.
The shortage may have also opened a very small window for Chinese firms to get in on the action. Shenyang’s Siasun Robot & Automation, for example, has started winning orders – although Ge of Nomura noted that Siasun mostly assembles imported high-value components into completed robots, a lower margin activity. Siasun declined requests for an interview.
But experts differ as to how easy it will be for new Chinese firms to enter the market or move up the value chain. So far industrial robotics remains dominated by a few big names from Germany and Japan: FANUC, ABB, Kuka, Yaskawa, and Kawasaki Heavy control over half the market, and have done so for decades. They are supported by giant component-builders like Siemens – which makes the delicate servomotors that drive robot arms – and other auxiliary systems providers that build control modules and user interfaces. Even Microsoft has a minor presence: Kuka uses a Windows XP interface for its internal operating system.
In some ways it seems odd that Chinese companies have not moved to attack the market. “There’s nothing stopping China from coming in,” said Alex Zelinsky, vice president at the Institute of Electrical and Electronics Engineers (IEEE) Robotics and Automation Society. “Industrial robots use very well-known technology. Most of the patents have expired.”
In fact, the field of industrial robotics is not considered to be a particularly high-tech industry. Most of the high-tech gee-whiz gimmickry associated with robotics is related to service robots, far different beasts. Industrial robots are still mostly built around 1980s technology. “It’s quite an old-fashioned business,” said Kirkwood of Global Robots. “The robot industry grew up to serve the car industry, and that industry paid for the development of industrial robots. So they are kind of stuck in an older age.”
Indeed, there are plenty of low-end robot components that are already being outsourced to China: chassis, controllers and the like. “I’m amazed that China hasn’t copied [the components],” said Kirkwood. While Ge of Nomura suggested that software and operating systems are where most of the value of a robot is captured, both Kirkwood and Zelinsky of IEEE said industrial robotics software development does not constitute much of a technology barrier. “If you can make a computer as good as what we had 20 years ago, you can build a robot,” said Kirkwood.
So why aren’t Chinese manufacturers moving more aggressively into the business? There are several issues holding them back. First, robot makers and their customers have sticky relationships. Robotics companies still use a vertically integrated model where they design both the hardware and the operating systems, while outsourcing lower-value components – similar to Apple’s integrated hardware/software model. This is not particularly efficient as far as customers are concerned, but it means that factory lines and factory workers tend to standardize around one brand
of robot for a particular application. Once staff is trained, there’s little reason to switch, especially since the management processes also need to be customized around the particular model.
Second, robotics is not a particularly high-margin industry, especially given the high level of customization most customers require. Ge said that the highest margin robot makers can earn is 10% – and for mere assemblers like Siasun, margins are likely even lower. Other forms of industrial automation, including numerically controlled machine tools and pneumatics, have higher margins and have attracted more Chinese competition – and most of the government support. But given that Chinese entrants don’t have much of a cost advantage when it comes to making industrial robots, there’s little incentive to attack a hoary monopoly running (mostly) on 1980s technology.
Third, of course, is the fact that Chinese robot makers don’t have much in the way of cost advantages. “The Chinese robots we have seen have been no cheaper than anybody else’s,” said Kirkwood. “There’s no point in buying them because they are an unknown quantity.” And China’s reputation for making cheap products that wear out fast doesn’t help in a market where customers want everything to be over-engineered.
Regardless, it looks to be a good decade for robot makers, be they foreign or Chinese. And the fact that robot makers are relocating production to China means they’ll be hiring Chinese workers. Win-win?
Not necessarily. The robot factories aren’t particularly labor-intensive; it’s doubtful they will directly
create as many jobs as they displace. And there are questions about the effect the transition will have on unskilled factory laborers.
In early August, Terry Gou, head of Taiwan-based electronics manufacturer Foxconn (the trade name of Honhai Precision Industries), announced that his company would add 1 million robots to its production floors – a number equivalent to the company’s current human headcount – by 2013. Given Foxconn’s high-profile problems with human resources management, most notably a cluster of employee suicides at Foxconn dormitories last year which embarrassed management, the press leapt to the conclusion that Foxconn was planning to get out of the business of hiring human beings: “Foxconn solves suicide problem with robots!”
Foxconn employees were also nonplussed. An engineer who recently left the company pointed out that adding 1 million robots in two years almost guarantees a reduction in headcount. To make use of them, he said, the company has to either increase production or lay off staff, but the increase in production would have to be massive to account for all those new robots. “Staff at the production line will worry about their job safety,” he said.
The announcement inarguably could have been handled better. Alexandra Harney, author of “The China Price” and expert on Chinese labor issues, found Gou’s choice of venue for the announcement darkly amusing: “Only at Foxconn would it be considered a good move to announce the replacement of workers with robots at a company employee party.”
Foxconn, unsurprisingly perhaps, denies any layoffs are imminent. A Foxconn human resources manager in the production division said that the robots are to be used to handle new production lines. The new robotic arms, which were designed and built internally, aren’t capable of displacing employees yet anyway, she said. “The robot arms’ production speed is now slower than sophisticated workers’ speed. Right now, a worker can operate three to eight machines while our robot arms can only handle two.”
Despite these conflicts, there is unlikely to be a major clash between labor and robots like there was in the US auto industry, where labor unions fought off automation for years. For one thing, there are fewer and fewer workers to lay off. China’s work force is projected to peak by 2015; Harney pointed out that the population of 15-19 year-olds, the typical shock troops of manufacturing lines, peaked back in 2005. Thus she doesn’t think robots are likely to do much net damage to labor’s bargaining position in China. “I can’t see how [robotics] would put downward pressure on wages given that the total supply of laborers is dwindling,” she said. “The reality is these factories are all scrambling to get enough workers.”
And there are other reasons to buy robots than to cut labor costs, some of which are in labor’s interests. Mainland manufacturers that responded to questions from China Economic Review mentioned increasing safety standards and improving precision and quality as their primary reasons to buy robots. Great Wall Auto, for example, plans to add 100 robots per year for the next three years, but said they will primarily be used for dangerous or monotonous tasks Chinese workers no longer want to do; they plan to move human roles off the production lines and add them in monitoring and maintenance positions.
Changan Automobile said it plans to invest “tens of millions of renminbi” in the next five years buying robots to use in welding and paint-spraying applications. The Changan public relations department pointed out that both of these applications entail long-term health risks for humans including respiratory diseases and leukemia.
Defendants of robotics do not deny that automation – and by extension technology in general – destroys some kinds of jobs, they just point out these jobs are usually worth destroying.
“The tractor put a lot of people out of work in agriculture,” pointed out Zelinsky of IEEE. “I think [the idea that robots take human jobs] takes a fairly narrow view of what productivity means.”
Proponents of robotics argue that by making companies more productive, robots create other, better jobs.
Yin Mingshan, CEO of privately owned automobile maker Lifan Group, concurred. “Increasing production capacity is the way to increase jobs in the robotics era. We have increased headcount since we adopted robots, because we are expanding our business scale all the time.” Yin plans to increase the number of robotic arms in use in his factories (currently at 16) to 100 over the next decade.
It is true that the jobs industrial robots will take are not the jobs Beijing particularly wants to preserve for its citizens. But there is a challenge nevertheless. It seems likely that some of the lower-skilled workers – especially men, widely considered by factory owners to be less reliable than women – will be swapped out to save on costs and reduce turnover.
“The higher the mechanization, the fewer the staff,” said Zhang Baojun, public relations director at Changan Automobile. He said the company plans to adopt “flexible labor policies to ensure fast company development.” And he allowed that using more robots would allow management to put more pressure on underperforming staff members.
But there are issues with implementation for both labor and management.
First, as the Foxconn case implies, just because the population of laborers is set to start shrinking does not eliminate the risk of layoffs. If Chinese manufacturers automate faster than the population declines, China could still end up with an unemployment problem at the bottom of the workforce.
Second, who are these employees who can repair robots, reprogram robots, design new manufacturing systems around robots? It doesn’t follow that the guy who used to weld car doors together will be transferred to the robot maintenance department. Automation will put pressure on China to adapt its education system to suit a more sophisticated manufacturing environment.
There are challenges for managers too. Buying robots is easy, but as Chinese factories automate, they become just like foreign factories, operating at similar price points; Chi
nese robotic production lines competing against foreign robotic production lines. Nor are robots able to fix flawed processes. General Motors tried and failed to fight off Toyota by using robots in the 1980s. It built some fancy robotic production lines that required a lot of troubleshooting, but skipped the sort of process streamlining that made Toyota so efficient. Critics said the company had only “automated confusion.”
But that doesn’t mean automating is optional. China needs only glance at the ashes of Detroit’s economy to see where rebellion against machines will end up. The robots are coming. Resistance is futile.