Tracking value-added trade : Examining global inputs to exports

by P. Cross and Z. Ghanem*

Introduction

In 1999, the CEO published the first study of the import content of exports in Canada.¹ This was updated in 2002.² These papers showed that firms in Canada were re-organizing their production processes in the 1990s to use imports more extensively. Manufacturers led this change, especially autos and the high-tech sector, but the shift was pervasive. However, the increased use of imports began to slow late in the 1990s, partly as the low exchange rate raised the cost of imports.

The heightened use of imports has significant implications. They are one barometer of firms increasingly outsourcing part of their production to lower costs.³ This globalization of production also altered how the Canadian economy performed over the business cycle: for the first time ever, Canada’s imports rose during a recession (in the early 1990s), while the impact of the 2001 recession in the US on our exports was absorbed more by a drop in imports and not lower output in Canada.

Comparing gross exports to GDP has always resulted in misleading analysis. Exports are the equivalent of gross sales, while GDP is measured on a value-added basis.⁴ Removing the import content of exports puts them on the same value-added basis, revealing the true exposure of GDP to external demand.

This paper extends the results from 1999 to 2004. This period covers several notable structural changes in the economy, including the meltdown in the ICT sector (culminating in the US recession in 2001) and the beginning of the dramatic rise of the Canadian dollar as the commodity price boom started late in 2002. The results will show how firms in Canada have begun to respond to these events. As well, this paper presents new data on imports of intermediate inputs versus finished products, including their origin by country.

The import content of exports can be calculated using the Input/Output (I/O) tables, which provide a comprehensive measure of the purchases all industries in the economy make in producing their products. By subtracting purchases from other industries from an industry’s gross output, we can derive value-added by industry as well as imports. Comparing imports with gross output allows us to isolate changes in import intensity by industry over time. Removing the import content of exports excludes both the direct content (such as an imported auto part in a vehicle export) and indirect content (imported steel that ultimately ended up in the same vehicle export).

To facilitate the analysis, we sometimes compress 59 industries into 21 major groups that summarize our main exports. As a result, large exporters such as autos and machinery and equipment are studied individually, while many services that export relatively little are grouped together (such as business services and finance). Industries that service the domestic market almost exclusively, such as construction and government, are in the other goods and services group.

The overall share of imports in exports, and hence value-added exports, are available from 1986 to 2004. Because of the introduction of the new industry classification system (NAICS), the detailed industry data are available on a consistent basis only from 1997. The 2004 I/O tables were only recently released, reflecting the complexity and fine level of detail of these calculations.

Why firms import inputs

The reasons behind purchasing inputs abroad are well-understood. In the words of Alan Greenspan, former chairman of the Federal Reserve, “Much of what is assembled in final salable form in one country increasingly consists of components from many continents. Being able to seek out the most competitive sources of labor and material inputs worldwide rather than just nationwide not only reduces costs and price inflation but also raises the ratio of the value of outputs to inputs—the broadest measure of productivity and a useful proxy for standards of living”.⁵

Almost all countries are using more inputs purchased abroad. One of the best examples is China, which imports many parts made elsewhere in southeast Asia or Japan. These components are then assembled, using China’s ample supply of cheap labour, before being exported.⁶

One clear implication is that exported from China does not mean ‘made in China’. In fact, some products are better described as ‘assembled in China’ from components made in Japan, Taiwan and Singapore. The result: “a finer division of labour is being established as national borders become increasingly porous with regard to organization of the production process and the Internet, modern international banking and more and more efficient transportation shrink the distance between countries”.⁷

Import content fell steadily into 2003

After peaking at 31.6% in 1998 and 1999, the overall import content of Canada’s exports fell every year to a low of 27.1% in 2003. This steady retreat in import content is the opposite of the increases earlier our studies had found for the 1990s.

Our 2002 study found that the slowdown in the use of imports in the late 1990s was concentrated in industries that had an above-average use of imports, notably autos and machinery and equipment. Conversely, industries with a relatively low import-intensity continued to increase their recourse to imports, implying industries overall were converging to the mean.

The recent drop in the import content of exports has been pervasive. From 2000 to 2004, 48 of the 59 industries reduced their import content of exports, while two were unchanged and only 9 increased. The drop in output in industries such as autos and ICT goods, where imports account for nearly half of exports, led the overall decline. But a majority of industries lowered their use of imports to produce exports, even as talk of globalization and offshoring mounted. This widespread decline in the use of imported inputs coincided with the dollar hitting a record low in 2002, the last year that the price of imports rose.⁸

The exchange rate affects the import content of exports (M/X) with long lags (Figure 1). The rise of the dollar in the late 1980s did not encourage more import use until the early 1990s, while its depreciation in the 1990s did not help slow import use until the second half of the 1990s and reduce it after 1998. The recent rally in the dollar only just began to encourage more import intensity in 2004, with the prospect of further gains as the dollar soared to parity.

Figure 1

Almost half the drop in import content after 1999 originated in autos and machinery and equipment, with declines of 3 and 4 points respectively. Nearly half of these industries’ exports are made from imports. The relatively large drops in exports by these high-import industries explain why the overall drop of 5 percentage points in the import content of all exports was greater than the drop for any of the individual industries.

Of particular note is the lower import content of auto exports. Despite a small increase between 2003 and 2004, this industry’s use of imports has been on a downward trend from its peak in the 1990s. This drop accompanied a sharp increase in the output coming from Asian-owned factories in Canada. This supports the findings of another CEO article that Asian-owned factories are no more likely to use imported parts (from the US or Asia) than the traditional American -based producers⁹.

Industries with a low import content in their exports in 1999 generally reduced their use of imports by about a point. Of the 3 (out of 21) industry groups that boosted their import intensity, only petroleum refining had an above-average use of imports, rising from 29% to 31%. The other two (tobacco and transportation) have a low import content.

At the same time that export demand slowed for our most import-intensive goods, exports shifted to natural resources such as energy and mining. These goods have a relatively low import content, which depressed overall import content as their share of exports rose. Several factors explain this low import use: natural resources tend to be relatively bulky, making their transport over long distances costly, and they are often produced in remote areas. As well, their production process is not as standardized as for many manufactured goods.

Figure 2

Import content rises with the loonie

The rapid appreciation of the Canadian dollar after 2002 appears to be slowly inducing firms to use more imports. The loonie rose from an average of US 63.7 cents in 2002 to 76.8 cents in 2004 (one-third of its eventual climb to parity with the US greenback late in 2007). While only 13 of 59 industries raised the import content of exports between 2000 and 2002, 15 did between in 2003 and 20 in 2004 as the dollar turned up. As a result, the overall import content of exports edged up from 27.1% in 2003 to 27.3% in 2004, its first rise in current dollars since 1998. The volume of imported inputs rose even faster as the rising dollar lowered import prices.

Of the 19 industries that boosted their use of imported inputs to make exports after 2002, it is probably no coincidence that a majority were in manufacturing (8) or transportation (2). These sectors faced the strongest competitive or cost pressures from the rising loonie and soaring oil prices starting late in 2002. These industries include transportation equipment, furniture, primary metals, metal fabricating, leather, chemicals, plastics and rubber, electrical equipment, trucking and sightseeing.¹⁰

Of the other 9 industries that raised import use, at least three also were adversely affected by the rising dollar, including the motion picture industry¹¹, forestry and services for forestry and farming. Two others were related to health care and hospitals, where rising costs were squeezing budgets. Overall, most industries using more imported inputs after 2002 faced the largest need to search out lower costs and improve efficiency.

Not all manufacturers facing intense pressure from the exchange rate (among other factors) increased their use of imports: this group includes textiles and clothing. Textiles and clothing have traditionally been among the leaders in using imported inputs to lower the cost of production. Their sharp drop in import use after 2002 may reflect the closing of numerous mills and factories, sometimes as production was shifted completely to overseas countries. In these cases, the demand for intermediate inputs would have dropped to zero.

Exports’ share of GDP declined

The share of gross exports in nominal GDP hit a record high of 45.6% in 2000, at the peak of the high-tech boom. Since then, gross exports fell to 38.4% of GDP in 2004. This drop of 7.2 points was split equally between a decline in the share of value-added exports in GDP and a falling import content of exports.

Subtracting out the import content of exports leaves the share of value-added exports in GDP. This shows that 27.9% of GDP came from exports in 2004. This was down from its peak of 31.4% in 2000, and close to its recent low in 1997. It is well below the often-quoted but misleading share of gross exports in GDP, which peaked at 46% in 2000 before settling at about 38% in 2003 and 2004.

Since 2004, gross exports as a share of GDP fell another 1.9 points to 36.3% in 2006, largely due to lower prices. This suggests that the share of value-added exports in GDP is about 27%, assuming the drop in gross exports continued to be evenly split between lower imports and value-added exports. With exports to the US currently accounting for 75% of all Canada’s exports, this implies that about 20% of our output is exposed to the risk from the slowdown in US growth. The share of jobs would be even lower, as exports remain a sector with above-average output-per-worker.

Figure 3

As the share of exports in GDP has fallen, exports have been less of a destabilizing influence in GDP after 1999 than before. The ratio of the standard deviation of annual real export growth to real GDP growth was 1.4 (3.1 versus 2.2) in the 1990s (a ratio of 1.0 would reflect perfect synchronisation between changes in exports and GDP).¹² However, this ratio rose to 3.5 between 2000 and 2007, as the standard deviation of exports rose to 3.5 while GDP’s fell to just 1.0: as export volatility increased, the volatility of output fell sharply.

Table 1 Exports in GDP

The increasing gap between the variability of exports and GDP reflects several factors, two of which are highlighted in this paper. First, the share of exports in GDP has been falling. And second, imports account for about one-third of exports, cushioning the impact of fluctuations in exports on domestic output. Other factors include significantly stronger domestic spending as the commodity boom boosted incomes in recent years, while better inventory control has helped stabilize real GDP.

Imported inputs and productivity

Changes in the use of imports by exporters broadly parallel changes in productivity in Canada. The increase in the use of imported parts in the 1990s accompanied rapid growth in labour productivity of 1.9% a year as firms restructured in the face of recession and free trade. Since 2000, the use of imports declined, and annual productivity growth has averaged 1.1% (its long-term average increase since 1961 was 2.1%).

This points to the falling use of imports in producing exports as one of many factors that help explain the recent slump in productivity that puzzled analysts and policy-makers early in 2007. Last year, we examined this slump in more detail, suggesting it was driven by labour shortages and the rapid shift of labour and capital to the booming resource sector.¹³ The results in this paper suggest that the reduced willingness of firms in Canada to use imports in their production process also may have contributed to the slowdown in productivity growth after 2000.

Imports for intermediate and final use

The drop in imports destined to be embedded in exports also contributed to a shift in the composition of Canada’s imports. Overall, imports destined to be intermediate inputs into production fell from 21.1% of all imports in 2000 to 17.8% in 2004. Most of this decrease originated in imports¹⁴ needed to produce exports, which fell from 12.1% to 9.1% in 2004. This reflected both the lower import content of exports and a slight drop in exports in absolute terms over that period. Meanwhile the share of imports used as inputs for production for the domestic market was steady at about 9%.

Figure 4

Instead, the most rapidly growing part of imports has been those destined directly for purchase by consumers and businesses (the latter was mostly investment goods). These accounted for 82.2% of goods imports in 2004, up from 78.9% in 2000. This may reflect the fact that for some products (notably electronics, appliances and clothing) production increasingly moved offshore. Their global supply chain would then be confined within Asian countries, which ship the finished product directly to markets in the developed world.

Geographic origin of imports

The drop since 1999 in the use of imported inputs tracks the broad geographic origin of imports into Canada. Imports from the US peaked at 68.2% of all Canada’s imports in 1998. Since then, imports from the US have risen only 1.7% a year on average, while imports from the rest of the world have grown an average of 8.7% a year. As a result, the US share of our imports has fallen to 54.6%, while overseas imports have risen from 31.8% to 45.4% (led of course by China, which by itself accounted for half the increase).¹⁵

Figure 5

The US was the source for 62% of intermediate imports into Canada. The US dominance reflects the need for parts imported to feed domestic assembly lines to be close by, especially as manufacturers adopt just-in-time inventory systems. Very few could wait for the month-long voyage from China, or support the cost of air travel from Asia or Europe.

The slight drop in the US share from 70% of inter­mediate imports in 1999 was concentrated in a few industries, notably autos, computer and electronics, clothing and mining. These reflect specific factors, including the increase in output by Asian-owned auto firms, the severe drop in Canada’s output of clothing and ICT goods (some of which also were sent offshore) and the increased importing of metal ores from Latin America for processing in Canada.

Interestingly, intermediate inputs account for two-thirds of all inputs from the US, versus less than 60% from the rest of the world. Excluding crude oil, the latter would be only about 50%.

Moreover, China’s exports consist largely of finished goods, not parts of intermediate inputs. In fact, until recently China itself was a model of a trading nation that imports many parts, assembles them and then exports the resulting products directly for purchase by consumers and businesses in North America and other markets.¹⁶ The shift in Canada’s imports from the US to Asia is consistent with a lower reliance on imported parts, and a higher rate of imports of finished goods.

Conclusion

This paper has examined the recent trend in the use of imported inputs by industries in Canada. It found that the sharp increase in imported inputs in the early 1990s began to slow later in the 1990s, hitting a peak in 1999. Since then, industries in Canada reduced their use of imported inputs until 2004. While the abrupt shrinkage of the ICT sector after 2000 and the downsizing of the auto sector led this decline, the drop in the use of imported inputs was pervasive. There was some evidence that firms were again beginning to increase their use of imported inputs after the loonie began rising in 2003, especially in the manufacturing sector.

The lower use of imported inputs sheds light on several widely discussed trends. It contradicts fears of a widespread offshoring of domestic production as firms adopt global supply chains. It suggests that firms in Canada have ample room to import more inputs as the soaring loonie increased competitive pressures, something they began to do in 2004. Finally, removing the import content from exports reveals Canada’s true exposure to export demand, an important piece of knowledge as analysts debate whether other countries can ‘decouple’ from the current slowdown in the US economy. Already since 2000, changes in exports had less of an influence on the course of GDP.

Recent feature articles

Notes