Descriptive Analysis of FDI

We have collected available data from Japan, the PRC, and the US to assess the current state of FDI linkages across the three countries. The FDI data is organized first at the country level and then at the industry level. After examining the FDI data, we analyze data collected by the American and Japanese governments on the overseas activities of foreigninvested firms. This data includes industry-level employment and sales data for the multinational firms.

3.1 Country Analysis

We start the country-level FDI analysis with bilateral FDI data that allows us to examine the relative importance and trends of the three target countries in each other's FDI portfolios. Table 1 [ PDF 12KB | 1 page ] shows US FDI stocks abroad from 1982–2007 in the top row, followed by FDI shares by region or country destination in subsequent rows. The last three columns in the table show the average annual growth rates in FDI stocks by destination. By region, Canada has seen a major decline in its role as a destination for US FDI, with 20.9% in 1982 but only 9.2% in 2007. Europe experienced the largest percentage gains in US FDI stock shares, from 44.5% in 1982 to 55.6% in 2007, while Latin America and other Western Hemisphere countries increased from 13.6% to 16.9% and Asia and the Pacific rose from 13.6% to 16.3%, which were similarly modest increases. The PRC's share of US FDI stocks has grown from below 0.1% to 1.0%, while Hong Kong, China's share has increased from 1.4% to 1.7%. Combining the PRC's share of US FDI stocks with Hong Kong, China's share produces a 2.7% share total in 2007, which puts the PRC above four of the European countries listed but below six European countries, Canada, and Japan as a host of US FDI. The leading hosts of US FDI stocks were the United Kingdom with 14.3% and the Netherlands with 13.3% in 2007.

As shown in the last three columns of Table 1, the PRC had the fastest growth for US FDI stocks in the 1980s and 1990s, with average annual growth rates of 28% and 41%, respectively. In part, this reflects the very low initial FDI stocks in 1982 (49 million USD) and in 1990 (354 million USD), the smallest of any of the individual countries listed. From 2000 to 2007, however, three other countries and one region recorded faster average annual growth in US FDI stocks than the PRC's growth of 14%. These countries and region are Austria at 32%, Luxembourg at 22%, the Netherlands at 18%, and the Middle East at 15%.

Data from the Bank of Japan provides Japan's FDI stocks abroad from 1996–2007, as shown in the top row of Table 2 [ PDF 12.6KB | 1 page ]. Subsequent rows show the shares of Japan's FDI stocks by region or country of destination. By region, Asia's share of Japan's FDI stocks has declined from 30.6% in 1996 to 24.3% in 2007, while North America's share has declined from 37.8% to 33.6%. Western Europe has recorded the strongest share gains, from 18.4% to 26.7%, followed by Central and South America with a share increase from 4.6% to 10.0%. The Asia regional losses are concentrated mainly in Hong Kong, China and the ASEAN countries of Indonesia, Singapore, and Thailand. In contrast, the PRC's share of Japan's FDI stocks has more than doubled over the 11-year period from 3.1% to 6.9%. The only other country with a larger percentage increase in Japan's FDI stock shares is the Netherlands, with a jump from 3.3% to 11.7%. By 2007, Hong Kong, China and the PRC combined hosted 9.1% of Japan's total FDI stock abroad. Only the US (31.9%) and the Netherlands (11.7%) hosted larger shares of Japan's FDI.

The last two columns of Table 2 [ PDF 12.6KB | 1 page ] show the average annual growth rates of Japan's FDI stocks abroad in the late 1990s and 2000s. The PRC received most of its increase in FDI stocks from Japan during the 2000s, with average annual growth of 23%, just above Thailand's growth rate of Japanese FDI stock of 22%. The growth rate for Thailand comes after a period from 1996– 2000 of FDI losses of 26% on average annually, probably due to the Asian financial crisis. Japan's FDI stocks in Hong Kong, China experienced a less severe decline of 8.9% on average annually during the late 1990s, then a similar average annual increase during the most recent seven-year period.

Having examined the FDI data from the perspective of the investing countries, we now turn to the host country perspective. The data for Table 3 is published by the PRC's Ministry of Commerce and is available for FDI inflows by investment source for a few major investors.⁴ As seen in Table 3 [ PDF 9.2KB | 1 page ], the growth in FDI inflows into the PRC was particularly rapid during the 1990s, with average annual growth rates of almost 28%. Branstetter and Foley (2007) note that these statistics are heavily influenced by official restrictions on direct investments from Taipei,China that prompt investment routing through Hong Kong, China or tax havens such as the Cayman Islands, the preference among some advanced country investors to invest in the PRC through Hong Kong, China-based affiliates, and the likelihood that many other investments purportedly from Hong Kong, China are actually PRC investors seeking to qualify for preferential treatment offered to multinational enterprises. For these reasons, interpreting the statistics at the individual investor country level is somewhat suspect. The investment shares by country shown in Table 3 indicate that both Japan and the US have decreased in relative importance as investors in the PRC, with Japan's share falling from 11.7% to 6.6% between 1986 and 2006 and the US share falling from 14.5% to 9.4% during the same time period. However, for the reasons mentioned above, these shares may exclude other investment that occurred through Hong Kong,China or tax havens. From the PRC's perspective, the Japan and the US are behind only Hong Kong,China and the EU as reported sources of FDI inflows into the PRC.

In sum, the PRC has seen very strong growth in its FDI stocks from Japan and the US in recent years, but the growth has been from low initial values so the PRC still hosts only small portions of their worldwide investments. Both investing countries have larger stocks of FDI in each other than they have in the PRC as of 2007. Of the US FDI stocks in 2007, 2.8% was invested in the PRC, including Hong Kong, China while 3.6% was invested in Japan. For Japan's FDI stocks, an enormous 31.9% was invested in the US and only 9.1% was invested in the PRC, including Hong Kong, China. The PRC plays a larger role in Japan's FDI abroad than in the US FDI abroad, probably due in part to its closer proximity to Japan. The breakdown of FDI by industry may help explain these different patterns of investment, a topic covered in the following section.

3.2 Industry Analysis

Table 4.1 [ PDF 12.6KB | 1 page ] and Table 4.2 [ PDF 12.4KB | 1 page ] show shares of US FDI flows by industry of affiliates for Japan and for the PRC for 1999–2007. The US FDI flows to Japan were dominated by the finance and insurance and information industries in 1999, with 44.0% and 22.6% shares, as shown in Table 4.1. By 2007, these industries contributed smaller shares of 21.0% and 7.7%, respectively, while holding companies contributed 59.1%.⁵ The contributions of manufacturing industry affiliates to US FDI flows to Japan has fluctuated widely over the nine years, with a high of 36.6% in 2000 and a low of -385.6% (i.e., capital inflows from affiliates) in 2003. The large capital inflows recorded from affiliates in other manufacturing, chemicals, and primary and fabricated metals, along with depository institutions, may have been caused in part by the yen's appreciation in 2003. Overall, non-manufacturing affiliates received the vast majority of US FDI flows to Japan over the 1999–2007 period.

In contrast, US FDI flows to the PRC have tended to favor manufacturing industries, as shown in Table 4.2 [ PDF 12.4KB | 1 page ]. The contribution has varied over the years with a high of 83.6% in 2001 and a low of 28.9% in 2005, but it rebounded to a 56.8% contribution in 2007. Among manufacturing industries, computers and electronic products and chemicals contributed the largest shares in 2007 with 25.9% and 12.9%, respectively. Among non-manufacturing industries, other industries (14.4%) and wholesale trade (13.2%) were the only industries with contributions above 6% in 2007.

Table 5.1 [ PDF 11.7KB | 1 page ] and Table 5.2 [ PDF 12.3KB | 1 page ] are comparable to Tables 4.1 and 4.2, but the focus shifts to Japan's FDI flows by industry for the PRC and the US for 1989–2004. As shown in Table 5.1, the manufacturing affiliates' share of Japan's FDI flows to the US varied over a relatively narrow range between 24% and 38%, except for a large jump to 60% in 2002 and 2003. This volatility reflects the influence of large investments made in particular years, specifically in electrical industries in 2002 and in chemical industries in 2003. Japanese investments in affiliates in the service, transportation, and real estate industries also appear to be particularly volatile from year to year.

Table 5.2 [ PDF 12.3KB | 1 page ] shows Japan's FDI flows to the PRC with shares by affiliate industry. Unsurprisingly, the recent shares for manufacturing industries are particularly high, 82.8% of total investments in 2004. This share has grown steadily over time, from only 47.0% in 1989. Among manufacturing industries, the transport sector had the largest FDI share in 2004 with 36.6%, up from 0.3% in 1989. The non-manufacturing industries have declined in importance as targets of FDI in the PRC, from 52.8% in 1989 to only 12.9% in 2004. The services sector saw the largest individual share decline, from 40.1% to 3.0% over the same time period.

The data in Tables 4.1–5.2 show more similarities than differences in Japanese and US investment flows. Both investing countries tend to make most of their investments in the PRC in manufacturing affiliates, while making most of their bilateral investments in each other in nonmanufacturing affiliates. One difference is that US investment in Japan is more concentrated in non-manufacturing than is Japanese investment in the US, while Japanese investment in the PRC is more concentrated in manufacturing than is US investment in the PRC. The industries that have consistently played a major role in Japan-US investment flows are finance and insurance for US investments and finance and insurance, service, and real estate for Japan's investments. For the PRC, US investment flows to computers and electronic products affiliates and chemical affiliates have been consistently high, while Japan's investment flows in the electrical, machinery, and transport sectors have been noteworthy.

The PRC's total FDI stocks and shares by industry for 1996–2007 are shown in Table 6 [ PDF 14.5KB | 1 page ].⁶ The National Bureau of Statistics of the PRC changed the industrial classification in its Statistical Yearbook from 2004, so we have divided the data into 1996–2003 and 2004–2007 periods. This division roughly matches with the PRC's pre-WTO and post-WTO entrance since the PRC officially joined the WTO on December 11 2001 and there were some lags in policy implementations.⁷ Manufacturing's share of FDI was 57.9% in 1996, increased fairly consistently to a high of 63.2% in 2006, then fell back to 58.3% in 2007. Unfortunately, the PRC's statistics do not allow us to break down the manufacturing sector into individual industries to see where changes may be occurring. Real estate had 19.4% of the PRC's FDI stocks in 1996, but its share fell to 13.4% by 2007. Social services had 6.6% of FDI in 1996, but this industry category was discontinued from 2004. In 2007, the only sector besides manufacturing and real estate with more than a 5% share of FDI was leasing and business services with 5.2%.

Of the 16 industry categories shown for the PRC's FDI stocks in Table 6, nine had faster average annual growth in the 2004–07 period than in the 1996–2003 period, while only three industries had slower growth (four industries had missing data that precludes a comparison across the two time periods). Of the three industries with slower growth in the 2004-07 period, two fell from rapid positive growth to negative growth, while one dropped from 36.5% to 28.3% average annual growth (scientific research, polytechnic services, water management, and environmental and public facilities). Transport, storage, post, and telecommunications service had 18.7% annual growth in the early period but -3.8% growth in the later period while “other had 29.1% annual growth followed by -37.2% growth. The industries with the fastest average annual growth rates were leasing and business services (81.1%), financial intermediation (69.9%), and information, computer services, and software (53.9%), all between 2004 and 2007. The first and last of these three industry categories were added from 2004. Two other industries with noteworthy increases in annual growth are wholesale and retail trade and catering services and real estate. Both jumped from 1% or less annual growth in the early period to 25% annual growth in the later period. FDI in mining also jumped from 2.1% to 20.2% average annual growth between the two time periods.

In sum, from the late 1990s up to the present, the majority of FDI stocks (57–63%) in the PRC have been in manufacturing industries, followed distantly by real estate with 11–19%. US FDI flows to the PRC have been less concentrated in manufacturing while Japan's FDI flows have been more concentrated in manufacturing, particularly in transport, electrical, and machinery industries in the most recent years. US investments in the PRC have included significant flows to wholesale trade industries and other service industries recently. To compare Japanese and US investments in the PRC in greater detail, we turn now to operating data collected directly from these countries' foreign affiliates.

3.3 Multinational Enterprise Activities Analysis

Both the US Bureau of Economic Analysis (BEA) and Japan's Ministry of Economy, Trade, and Industry (METI) conduct detailed surveys of their firms' international activities. One noteworthy difference across these two surveys is that the BEA survey is more comprehensive in its coverage since reporting is mandatory for participating firms while the METI survey is not mandatory and generally posts response rates between 60–80% of recipients. To improve the METI data, Matsuura (2004) uses the METI firm-level responses to create a panel data set and he estimates missing values in each year based on each firm's responses in other years. He then aggregates the responses to industry and country levels and posts the data on a website maintained by the Research Institute of Economy, Trade, and Industry (RIETI).⁸ The most recent updates to the RIETI website (June 2009) provide data on the employment and sales activities of Japan's Majority-Owned Non-Bank Affiliates (JMONA). This data can be directly compared with BEA data on US Majority-Owned Non-Bank Affiliates (USMONA).

Table 7 [ PDF 12.6KB | 1 page ] and Table 8 [ PDF 11.6KB | 1 page ] present employment statistics of USMONA and JMONA in the PRC for selected years between 1990 and 2005. USMONA employed only 13,600 workers in the PRC in 1990 but this number grew to 521,800 by 2005, as shown in Table 7. The growth in employment was particularly rapid in the 1990–95 period, with 43% average annual growth, followed by average annual growth rates of 25% in 1995–2000 and 16% in 2000–05. The industry breakdown indicates an early concentration in manufacturing—with 75% of employment in 1990—that became stronger in the mid-1990s—with 90% in 1995—but then has weakened in recent years to only 60%, as the shares of employment in non-manufacturing sectors have grown. At the individual industry level, USMONA in electrical equipment, appliances, and components employed 52% of the total USMONA workforce in the PRC in 1995. This industry grouping later broke out a separate computers and electronic products grouping which had 28% of the USMONA workforce in the PRC in 2000, while electrical equipment, appliances, and components employed 15%. The top individual industry employer in 2005 was other nonmanufacturing with 29%, followed by computers and electronic products with 21%.

The data for USMONA presented in Table 7 [ PDF 12.6KB | 1 page ] can be compared with the data from JMONA shown in Table 8. In the PRC in 1990, JMONA employed more than three times as many workers as USMONA. By 2005, JMONA had less than twice as many employees as USMONA (967,100 versus 521,800). The growth rates for JMONA employment in the PRC were particularly strong in the early 1990s, with 46% average annual growth. Interestingly, JMONA also increased their concentration in manufacturing between 1990 and 1995, from 80% to 92%, but unlike their US counterparts, the concentration in manufacturing employment remained very stable in the 92–94% range through 2005. In 1990 among individual industries, JMONA employed the most workers in other manufacturing (26%), computers and electronic products (19%), and electronic and other electric equipment (16%). Only one of these three industries, computers and electronic products, saw its share of JMONA employment in the PRC grow through 2005, to 30%, making it the largest industry in terms of employment in that year. The second largest industry in terms of employment in 2005 was transportation equipment with a 19% share, followed by other manufacturing with 12%. In the transportation equipment sector JMONA recorded a rapid average annual growth rate of 82% between 1990 and 1995, while the average annual growth rate for all manufacturing industries combined was 50% during that period. In manufacturing JMONA continued rapid employment growth in the PRC with average annual growth rates of 46% from 1995–2000 and 40% from 2000–2005.

The sales activities of USMONA and JMONA in the PRC can be compared using Table 9 [ PDF 12.9KB | 1 page ] and Table 10 [ PDF 9.8KB | 1 page ]. Table 9 shows the destination shares of USMONA in the PRC, with sales grouped by local sales, exports to the US, and exports to other countries for selected years from 1989–2005.⁹ At the all industries level, in the PRC USMONA made 94% of their sales locally in 1989, but the local sales share fell to 78% in 1994 and then to 73% in 2005. Much of the shift in sales favored exports to other countries, with a rise in sales share from 5% to 19%, but exports to the US also had a rising share from less than 1% up to 8% of total sales. Manufacturing industries in aggregate show a similar decline in local sales share, from 88% in 1989 to 67% in 2005 and a corresponding rise in exports. At the individual manufacturing industry level, however, there is significant variation in the pattern of sales over time. USMONA in primary and fabricated metals reported local sales of 100% in 1989, 75% in 1994, back up to 97% in 1999, down to 88% in 2002, and then 73% in 2005. This industry showed the largest percentage point decline in local sales between 1989 and 2005, followed by the machinery sector with a local sales share decline from 82% to 64% over this period. Electrical equipment, appliances, and components also show a declining pattern of local sales, from 56% in 1994 to 39% in 2005, with sales shifting to favor exports to the US in 1999 and 2002, but shifting to favor exports to other countries in 2005. Only two industries, computers and electronic products and transportation equipment showed increases in their local sales share between 1999 and 2005, from 44% to 59% and 74% to 84%, respectively.

Unsurprisingly, among the non-manufacturing industries shown in Table 9 [ PDF 12.9KB | 1 page ], we see quite high local sales ratios over the entire time period since most of these industries are services. Unfortunately, many of the observations for individual industries are either not available, due to changes in industry classification, or suppressed due to disclosure concerns with small numbers of affiliates reporting. One interesting change is the decline in local sales for professional, scientific, and technical services from 94% in 1994 to 67% in 2005. For this industry, sales shifted to strongly favor exports to the US, which increased from less than 1% of sales in 1990 up to 25% in 2005. However, USMONA in these services industries contribute only a small share of the total sales by USMONA in the PRC. The total sales for USMONA in the PRC in these services industries were 914 million USD in 2005, approximately 1% of total sales for all USMONA in the PRC.

Table 10 [ PDF 9.8KB | 1 page ] shows the distribution of sales by JMONA in the PRC for selected years between 1989 and 2005. The sales shares suggest that in 1989 in the PRC JMONA were more export-oriented than USMONA. Half of JMONA sales went to export markets while only 6% of USMONA sales were exported. By 2005, 44% of JMONA sales were exported versus 27% for USMONA. For JMONA, the destination of their exports changed to favor Japan, which increased its share from almost 10% in 1989 to 24% in 2005. The share of sales exported to other countries fell from 40% to 19% over the same time period. The share of JMONA sales going to the local PRC market increased slightly, from 50% in 1989 to 56% in 2005. Comparing the manufacturing industry totals across JMONA and USMONA shows a similar pattern of very divergent patterns in 1989 but some convergence over time. In manufacturing in 1989 JMONA were exportoriented, with 53% of their sales exported, while USMONA were strongly oriented towards the local market, with only 12% of their sales exported. Over time, in manufacturing JMONA become somewhat less export-oriented (46% of sales exported in 2005) while USMONA become more export-oriented (33% of sales exported in 2005).

Looking at the individual industries in Table 10 shows significant variations over time. Among the manufacturing industries, electronic and other electrical equipment and computers and electronic equipment are the most important industries in terms of sales. Their sales pattern is extremely export-oriented initially but becomes less export-oriented over time. Electronic and other electrical equipment exported 79% of their sales in 1989 and computers and electronic equipment exported 89%, but these export numbers drop to 52% and 67%, respectively in 2005 as local sales grew, from 20.6% to 47.0% for electronic and other electrical equipment and from 11.1% to 33.0% for computers and electronic equipment. Among exports, the electronic and other electrical equipment industry maintained a fairly stable share of sales exported to Japan—40% in 1989 and 36% in 2005—while the share exported to other countries dropped from 40% to 16%. In the computers and electronic equipment industry, exports to Japan grew—from 3% in 1989 to 33% in 2005—while exports to other countries declined precipitously, from 85% to 34% in the same time period. Some of the other manufacturing industries showed the opposite trend, moving from selling the vast majority of their products in the local market to exporting significant shares.

The sales shares in Tables 9 and 10 indicate that in the PRC JMONA had a much stronger export-orientation relative to USMONA in 1989, but this gap has lessened in more recent years. Interestingly, Lipsey (2000) found a similar trend towards convergence but from opposite starting points when he examined Japanese and US manufacturing affiliates in other East Asian economies using data from 1977–1995. He found that US affiliates in East Asia tended to be more export-oriented than their Japanese counterparts in 1977. The US affiliates on average became slightly less exported-oriented between 1977 and 1995—with export shares of total sales falling from 57% to 54%—while Japanese affiliates became more export-oriented—with export shares that grew from 33% to 44%. The differences between our results and Lipsey's (2000) results suggest that Japanese and US multinationals have approached their investments in the PRC differently than they approached investments in other East Asian countries previously. At the individual industry and country level, Lipsey (2000) noted a lot of variance in export shares between 1977 and 1995 and more recent data on JMONA and USMONA activities in other East Asian economies confirms significant variance in export share levels and time trends.¹⁰ Further comparisons of JMONA and USMONA activities in other East Asian economies versus their activities in the PRC will be deferred to future research.

Download this Paper [ PDF 138.1KB| 33 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.