Keith Bradsher has an interesting piece in the New York Times on GM's recent success producing and selling cars in China. The interesting fact is the way in which China's relative abundance of labor altered GM's capital investment:

In this obscure corner of southern China, General Motors seems to have hit on a hot new formula: $5,000 minivans that get 43 miles to the gallon in city driving. That combination of advantages has captivated Chinese buyers, propelling G.M. into the leading spot in this nascent car market....

To build the cars, G.M. helped gut and rebuild a former tractor factory in ways that could become a model for automobile production in China for years to come.

Long white halls erected in 1958 during Mao's Great Leap Forward still stand here, the paint peeling in places, the wood window frames warped and the windowpanes cracked and broken. Inside, however, is a factory that combines old and new management techniques. Small, plastic racks of parts delivered several times or more a day have replaced large bins of parts delivered to the assembly line in big shipments every few days. This way, the factory can keep low inventories and order quick design changes, if necessary, from nearby suppliers.

The assembly process has only one robot, for sealing windshields, relying mostly on workers earning $60 a month, above average for this impoverished region. That comes after G.M.'s experience in Shanghai, where it installed four dozen robots for its first assembly line only to find them much costlier and less flexible than people; G.M.'s second assembly line there was built with only four robots.

"Low cost doesn't just mean low wages, it means low investment," Mr. [Stephen] Small [the G.M.-appointed chief financial officer of the joint venture] said.

Worker safety in most Chinese factories is abysmal by Western standards. But workers at the factory here wear safety glasses, and the equipment has automatic cutoffs to prevent workers from losing fingers.

The depressing fact is that, naturally, GM is punishing the guy that came up with the process and product ideas behind the minivan in the first place:

Their development was led by an American, Philip F. Murtaugh, a native of Ohio and a maverick executive who was willing to zig while the rest of G.M. was zagging. Mr. Murtaugh was able to create in China the kind of innovative environment that G.M. has struggled for decades to achieve in its American operations. But whether G.M. can duplicate elsewhere its achievements in China or even keep its pace here is unclear.

In what may be a telling sign of the corporate culture at G.M., Mr. Murtaugh's success in China led not to promotion but to his departure from the company. G.M. declined to discuss personnel matters, but both it and Mr. Murtaugh said he resigned and was not dismissed.

A soft-spoken man in a company known for autocratic leaders, Mr. Murtaugh ran the China operations for more than nine years from his base in Shanghai, repeatedly making some of the best calls in the industry. Now he finds himself unemployed and living in a small community in rural Kentucky.

His resignation in March, at the age of 49, came shortly after senior company executives reorganized management to give more power to Detroit executives to oversee design, engineering and various manufacturing disciplines all over the world, including operations in China.

In a world where local knowledge about consumer demand and the most efficient way to mix factor endowments are important, the GM decision to centralize its management structure seems particularly brain-dead.

Read the whole piece.

UPDATE: A 2003 McKinsey Quarterly essay by Vivek Agrawal, Diana Farrell, and Jaana K. Remes touches on this concept of reorganizing production processes to exploit local factor endowments. The auto sector in China is one example:

Certain automotive original-equipment manufacturers (OEMs) in China use robots for only 30 percent of the welding in car assembly, as compared with 90 percent or more in US or European operations. (BMW’s plant in South Africa employs the same line of attack.) In India, domestic car companies have reduced the need for automation throughout the manufacturing process: they use more manual labor to load and change dies in pressing, body welding, materials handling, and other functions—while suffering no discernible loss of quality in the finished product. In this way, these companies manage to cut their assembly costs by 4 to 5 percent or even more and save themselves millions of dollars annually.

Ultimately, companies might completely redesign the sequence in which tasks are performed, in order to leverage the opportunities above more fully. Consider the simple example of a call-center agent who manages customer accounts. In high-wage countries, each customer call is routed to an agent who listens to the request, opens up a computer database, and updates the account in real time. Neither the computer nor the telephone is used efficiently, since the agent is either talking or typing but not both.

Offshore, an agent equipped with only a telephone could write the customer request by hand into a tracking log and move on to the next call. Telecom costs are reduced because the agent spends less time on calls and customers less time on hold. Another agent, working at a computer station used around the clock, could enter the information into the database. While the new process requires more agents to handle requests, expensive computer hardware and software and telephone lines are used more intensively. Added wages are more than offset by savings on computers, software licenses, and telephone connections (Exhibit 4). The economics of an Indian call center suggest that this simple change could actually boost current profit margins for offshoring vendors by as much as 50 percent.

These examples point to a big warning sign that should be put on any news story about job creation in offshored sectors in low wage countries:

WARNING: THESE ARE NOT JOBS THAT WOULD EXIST IN THE UNITED STATES. THESE TASKS WOULD BE AUTOMATED.