Exchange rates have been known to stray from their fundamentals for long periods of time, so companies need an adjustment strategy even if they believe the currency will eventually ease.

Textbooks tell us that a period of exchange rate overvaluation will see exports weaken and the economy slow, as domestic companies find it difficult to compete internationally. Profit margins are squeezed, so investment falters. Workers are laid off. Disinflationary pressures set in, leading to lower interest rates, the currency eases back and the economy picks up again.

As is usually the case, the real world is more complex and it is helpful to examine the U.S. experience with its overvalued currency. In early 1995, the U.S. dollar trade-weighted index against a basket of currencies was at about 80. It then rose continuously for six years, as the world staggered from one crisis to another and risk-averse investors sought the relative safety of American markets. The dollar index peaked at about 108 in early 2002, an appreciation of some 35 per cent. As the world economy recovered, the dollar eased back, and today it is back around 83.

How did American companies react to the overvalued dollar during the early years of this decade? First, they globalized their operations, offshoring segments of their business to lower-cost foreign markets.

One measure of this phenomenon is a country's trade penetration (exports plus imports as a share of GDP). In 1994, U.S. trade penetration was about 22 per cent, but during the next six years it rose by about five percentage points, a major increase.

The ratio has fluctuated during the past seven years, but lately is around 28 per cent. A related measure of globalization is American investment abroad, which also rose significantly during the rising dollar period.

Second, U.S. companies increased their spending on cost-saving equipment at home. Investment in new equipment and software as a share of GDP was pretty flat during the early 1990s, at five per cent. By 2000, this ratio hit 9.4 per cent of GDP, and has fluctuated in the eight- to nine-per-cent range in recent years.

Third, U.S. manufacturers reduced employment at home. Manufacturing employment was just over 17 million in early 1995, 14.8 per cent of total employment.

Manufacturers began to cut jobs steadily in early 1998 and at the dollar's peak in early 2002 employment was down to about 15.5 million, around 10 per cent below the early-1995 level.

The downtrend has continued since then - the most recent figures show 14 million manufacturing workers, 10.1 per cent of total employment.

Very importantly, though, U.S. manufacturing output has continued to grow throughout this period of employment compression. Between early 1995 and today, U.S. manufacturing employment has declined by nearly 20 per cent, whereas manufacturing value added (its contribution to U.S. GDP) is up by some 50 per cent over the same timeframe - implying a huge productivity improvement.

The bottom line? It is impossible to say whether Canadian companies will adjust to a strong currency in the same way, but it is reasonable to expect to see some of the same elements. The U.S. experience does show that large, successful adjustments are possible in the right conditions.

(Stephen Poloz is a senior vice-president and chief economist for Export Development Canada. He can be reached at spoloz@edc.ca)