As U.S. auto assembly lines grind to a halt for want of components that usually come from now-disabled factories in northeastern Japan, business strategists may be forced to rethink the way globalized companies do business.

The effects of Japan's March earthquake and tsunami are being felt far beyond the shattered region around Sendai and Fukushima. As U.S. auto assembly lines grind to a halt for want of components that usually come from now-disabled factories in northeastern Japan, business strategists may be forced to rethink the way globalized companies do business. The result could well be a retreat from current manufacturing methods -- sourcing key components from a single supplier and running "lean" factories without stocks of supplies on hand -- whose main goal is to minimize costs. Now, management may also pay close attention to risks.

Such a change would represent a reversal of course for major international companies, potentially transforming the way that many of the world's industrial giants have functioned for the past two decades. Whereas companies used to run separate operations in many countries, each serving a given national market, in the 1980s multinational corporations started to run their affairs with diminishing attention to national borders. Today, a single plant or research center will often take worldwide responsibility for a particular product or business area. And whereas factories once manufactured their own components or purchased them nearby, now even some small plants have supply lines that stretch across the globe. Almost every manufacturer, from your local maker of wedding dresses to Boeing and Caterpillar, is a global company, because its production relies critically on parts or other inputs made or designed outside its home country.

The globalization of manufacturing is responsible for much of the boom in international trade over the past two decades. Around half of the maritime shipping containers that arrive in the Los Angeles and New Jersey ports, for example, contain not products for retail sale but "intermediate goods," products partially manufactured in one location and destined for further processing somewhere else. Similarly, a large proportion of airfreight consists of high-value components, such as semiconductors and optical lasers, rather than finished consumer goods.

The shift to dispersed, highly specialized operations has significant economic advantages. One is that each producer focuses on the particular things it does best, instead of devoting capital and management attention to a wide variety of activities or to activities that offer poor returns, such as manufacturing components that are already widely available at low prices. A second advantage of this type of production is that it permits economies of scale. A specialized factory may produce only a few types of goods, but it is able to make and sell a lot of each one, since its market is global. This usually means lower costs to produce each unit.

Financially, specialized production and global supply chains make sense. But firms -- and the investors who own them -- have paid too little attention to the risks that such a strategy entails.

As I wrote in Freight Pain (Foreign Affairs, November/December 2008), the development of long supply chains was made possible by the decline in the cost of moving goods. But even in 2008, those costs were already rising: "Congested shipping lanes and highways make transit times uncertain," I wrote, "and this uncertainty hurts profits." Moreover, the push for ever-greater port security will further slow transit; physical inspection of shipping containers could delay delivery by two to three days or more. As I wrote then, "Even if the proportion of containers pulled out of the flow of traffic is small, importers will be forced to reckon with the possibility that their goods might be delayed in transit. In some instances, importers will adjust by keeping more stocks in their U.S. warehouses at any one time."

Even more costly, an unforeseen event at one end of the chain may disrupt the flow of components to the next firm thousands of miles away. The problem is compounded if a sole-source supplier stops production. A manufacturer may be left high and dry if its producer of fuel pumps, friction bearings, or graphics processors goes offline. The recent disaster in Japan is an extreme example, but supply- chain problems are far from unusual. A shortage of freight cars, a dock strike, a factory closure due to lack of water, a riot that keeps a supplier's employees from getting to work -- these and many other events can wreak havoc with production schedules.

It is hard to quantify how much supply-chain interruptions have cost business in recent years, because they generally go undisclosed. After all, no company wants to alert its competitors to weaknesses in its business model. And even in the case of the Japanese disaster, many companies with lower profiles than the automakers have yet to announce how they will be affected by shortages of or higher prices for essential components. Public attention has focused on outages at Japanese plants that turn out silicon wafers and memory chips, but there are surely many more obscure products that are in short supply.

The total cost of lost U.S. production due to shortfalls of Japanese components will easily run into the billions of dollars; slowing or closing down auto assembly lines, as several manufacturers have already done, does not come cheap. Other products that are "made in America," particularly factory equipment, also rely heavily on Japanese electronics, and even if those producers are not forced to close plants temporarily, the risk of delays in filling orders may cause impatient customers to defect to competitors who do not depend on inputs from Japan.

The supply-chain problems that followed the Japanese disaster were not inevitable. Companies can avoid them by building supply chains with redundancy and resilience. Redundancy means manufacturing key components in separate locations and transporting them along different paths so that a single problem at a supplier, port, or warehouse does not cripple production. Resilience means backing away from just-in-time production strategies and maintaining larger inventories so that delayed shipments do not shut down an entire factory.

Redundancy and resilience are not cheap. Sourcing from multiple locations will result in higher input costs. Keeping larger stocks on hand is costly, too, since it forces manufacturers to finance goods sitting uselessly on their warehouse shelves. All of this represents a drag on the larger economy insofar as it reduces economic efficiency.

But while reducing supply-chain risks has its costs, idle auto plants and electronics factories around the world demonstrate that there is also a cost to doing nothing. When the global supply chain fails to deliver a key component on time, the expense may be much higher than corporate strategists imagine.