These prognostications are too gloomy. Experience suggests that export revivals are led by highly productive industries that already export a lot, rather than less productive industries regaining old markets. Over time a favourable global environment encourages more firms to launch new products or enter new markets abroad. Yet pinpointing those firms in advance is almost impossible.

In a 2001 paper Andrew Bernard of Dartmouth College and Bradford Jensen, now of Georgetown University, analysed the sources of America's export boom in 1987-92. During that period American exports rose 77% in nominal terms and the trade deficit shrank from 3.2% to 0.6% of GDP. They looked at factories across America and found that only a small portion of the rise in exports came from those that had never exported before; by far the biggest gain was from those that were already exporting. Their research suggested that 90% of the export boom could be explained by the dollar's depreciation and by relatively stronger growth in America's trade partners.

A later paper, published in 2009, which Messrs Bernard and Jensen co-wrote with two other scholars, looked at the sources of export growth between 1993 and 2003. The increase in any given year came almost entirely from existing exporters, but over time new ones played an increasing part. For example, in 1993 alone firms exporting existing products to existing markets accounted for 91% of total export growth. But over the entire ten-year period they made up a relatively modest 35% of the total, with new firms contributing 24% and firms with new products or entering new markets as much as 42%.

Mr Bernard says these findings are at odds with the conventional view that export growth is the result of domestic firms becoming more productive. That is true only in the long run. In the short term almost all the growth comes from higher demand for products that existing firms already make. That fills him with optimism: “If you wanted to set the stage for US exporters to do well, you'd like to have rapid income growth in destinations that will be big in GDP terms. We already see that.”

America's export boom is likely to be led by firms that are already global in scale and by sectors in which America has a clear competitive advantage: sophisticated, knowledge-intensive capital goods like microprocessors, and high-end services like engineering, oil-production services and even (witness KPF) architecture.

Few companies better capture that trend than Intel, a microprocessor giant. It is one of America's most successful companies, and 80% of its revenue comes from outside its home country. Like almost all big companies it suffered a decline in sales and profits in the recession and had to lay off some staff. Yet it also decided to invest $7 billion in its semiconductor operations to enable them to make chips with circuitry just 32 nanometres (or billionths of a metre) wide, down from 45.

Brian Krzanich, Intel's vice-president of manufacturing, says labour costs are an important consideration in processes such as assembling and testing systems, which are carried out in Costa Rica, Vietnam, China and Malaysia. But they matter much less in capital-intensive processes such as fabrication where the value of a technician or engineer might depend on how efficiently he or she can operate a $70m tool. If a more productive engineer can get 2% more use out of that tool, “that's worth a lot of employees.” The difference in pay between a $50,000 engineer in China and a $150,000 engineer in America is a “nit” in that equation, he says. Thus the bulk of Intel's investment is going into American factories in New Mexico, Oregon and Arizona. By the fourth quarter of last year Intel started shipping the new, smaller generation of chips from its Oregon plant, which helped the company achieve its highest ever gross margin.

Still, over the medium term Intel cannot correct America's trade deficit by itself. More American companies will have to look abroad. They are the least likely to export out of 15 big economies, according to the National Association of Manufacturers, a trade group. Matthew Slaughter of Dartmouth College notes that only 4% of all American firms and 15% of American manufacturers do any exporting at all. And 80% of America's total trade is conducted by just 1% of firms that export or import.

This does not mean there is something wrong with American firms. Rather, it reflects the fact that America's domestic market is large enough for most firms. Exports as a share of GDP last year were estimated at 10.9%, much lower than in most of its big trading partners (see chart 5). Smaller companies are deterred by the investment in market research, distribution and product design needed to sell abroad. To succeed in foreign markets, they first have to do well at home.


The definition of exports has also become less clear-cut over the years. These days much of what America exports is buried deep inside products put together elsewhere. Apple's iPod, for instance, is assembled in China, but a study by the Personal Computing Industry Centre at the University of California, Irvine, estimated that China accounts for only 2% of the wages involved in its manufacture, whereas America makes up around 70%, in areas such as engineering, software and distribution.

Service with a smile

Services are playing an increasingly important part in America's exports. Their share of the total has gradually increased, to nearly 33% last year. Within that category, the private side, which covers things like business, professional and financial services, has been growing fastest. Before the recession optimists routinely cited America's pre-eminence in financial services as a driver of future export growth. Since then the rest of the world has become much more sceptical about the real value of America's financial engineers.

Yet America leads in many other kinds of services, from software and film to engineering and oil drilling. And as technology advances, more businesses are likely to turn global. A survey by the American Institute of Architects found that in 2008 some 7% of its members' billings came from international work, against only 2.8% a decade earlier. That increase came almost entirely from firms that were already doing international work but had stepped up their efforts.

Building up such work can take time. KPF, for example, though founded in 1976, did not land its first foreign job until the late 1980s. Getting a toehold in Asia was particularly time-consuming, but the firm now has an office with 40 people in Shanghai that helps with local project oversight, models, presentations to clients and the notoriously opaque building regulations in Chinese cities.

The ease of exchanging information around the world has fuelled fears that even service jobs will be outsourced away from America. But those information flows also make it easier for firms like KPF to do business overseas. Once a week Ms Cusumano and her colleagues in Shanghai log into a teleconference to discuss the Hangzhou hotel and look at drawings with the aid of web-based conference software. And some nights when Ms Cusumano cannot sleep, she exchanges e-mails with her partners in Shanghai.

From the print edition: Special report