FOREIGN DIRECT INVES TMENT IN CHINA:
Effects on Growth and Economic Performance
Edward M. Graham
Institute for International Economics
Institute for International Economics
To appear in Achieving High Growth:
Experience of Transitional Economies in East Asia ,
Peter Drysdale, editor, Oxford University Press, forthcoming 2001.
The authors wish to thank Dr. Chi Zhang (Stanford University), Dr. Jun Ma (Deutschebank),
Professor Lori Kletzer (University of California and IIE), and Dr. Aasim M. Husain (International
Monetary Fund) for valuable input and comments on earlier drafts of this paper.
By almost all accounts, foreign direct investment (FDI) in China has been one of the major success stories
of the past 10 years. Starting from a base of less than $19 billion in 1990, the stock of FDI in China rose
to over $300 billion at the end of 1999. Ranked by the stock of inward FDI, China thus has become the
leader among all developing nations and second among the APEC nations (only the United States holds a
larger stock of inward FDI).1 China’s FDI consists largely of greenfield investment, while inward FDI in
the United States by contrast has been generated more by takeover of existing enterprises than by new
establishment, a point developed later in this paper. The majority of FDI in China has originated from
elsewhere in developing Asia (i.e., not including Japan). Hong Kong, now a largely self-governing
“special autonomous region” of China itself, has been the largest source of record. The dominance of
Hong Kong, however, is somewhat illusory in that much FDI nominally from Hong Kong in reality is
from elsewhere. Some of what is listed as Hong Kong-source FDI in China is, in fact, investment by
domestic Chinese that is “round-tripped” through Hong Kong (see footnote 2). Other FDI in China listed
as Hong Kong in origin is in reality from various western nations and Taiwan that is placed into China via
Hong Kong intermediaries. Alas, no published records exist to indicate exactly how much FDI in China
that is nominally from Hong Kong is in fact attributable to other nations.
According to official sources, in the period 1992-96, FDI from developing Asian nations
dominated total FDI flows into China, but since 1996 a growing portion of these flows has come from
other sources (i.e., Europe, North America, and Japan). This latter FDI generally has been of a different
character than FDI from developing Asian nations. While the latter has been concentrated in export-
processing activities in sectors in which China has revealed comparative advantage, much of the former
has been directed more toward the domestic market in sectors in which China has no revealed
comparative advantage. Thus one consequence of a rising percentage of FDI from Japan, Europe, and
North America has been that overall the activities of foreign-invested enterprises2 in China have become
somewhat more focused on the domestic market, and less on export markets, in the la te 1990s relative to
the mid-1990s. The consequences are discussed in more detail later in this paper.
1 However, some of the stock is known to be the result of “round tripping”, i.e., Chinese domestic savings
intermediated as direct investment in order for the investment to receive incentives available to foreign investors that
are not available to domestic investors. Most of this investment is recorded as being from Hong Kong. The exact
extent of round tripping is not known, and it is not discussed further in this paper.
2 “Foreign invested enterprises” is a somewhat euphemistic term used in China to denote local affiliates of foreign-
owned firms. Many of these local affiliates are joint ventures with Chinese enterprises and, indeed, until 1992
almost all foreign direct investment in China was in the form of joint ventures. The term “foreign-invested” thus
seems to have been employed to reassure anyone who might care that these ventures were really domestic firms with
These changes notwithstanding, there is little doubt that FDI has contributed significantly to
Chinese economic development: much and perhaps most of the growth of China’s exports can be
attributed to foreign-invested enterprises and per capita income growth in those regions of China where
FDI is concentrated has been significantly higher than in other regions. Specifics of some of the major
empirical fin dings are discussed in the following two sections of this paper. Section 2 reviews some
findings published elsewhere, and section 3 presents some additional findings by the authors.
However, the story of FDI in China is not quite as rosy as these summary sentences suggest. By
all accounts, the policy environment for foreign direct investors in China is difficult, and much anecdotal
evidence suggests that some of these investors are becoming discouraged by this environment while other
potential investors have been deterred by it.3 This discouragement seems to be indicated by the fact that
the flow of FDI decelerated in the late 1990s. However, data for the most recent years are almost always
subject to significant future revisions, so it is premature to read too much into the 1999 or even the 1998
data. Keeping this in mind, the number of investments in 1998 declined relative to 1997 by a larger
percentage than FDI flow, indicating that the average size of the individual investment undertaking rose
somewhat. This rise in average size of investment was driven by the changing source of the investment;
Japanese and western direct investors tend to make larger investments than investors from developing
Asia. Also, it should be kept in mind investment from developing Asian sources was negatively affected
by the Asian financial crisis of 1997-98 and that the impact of the crisis was felt most heavily in 1998.
Anecdotal evidence suggests that it is western and Japanese multinationals that are most discouraged by
the current policy environment in China. Thus, if the Chinese are counting on FDI from the west and
Japan to continue to make up for declining FDI from non-Japanese Asian sources, there might be cause
In this light, it is slightly ominous that US and Japanese data suggest that FDI from these
countries to China has dropped substantially in 1999 over levels of 1998 (see tables 1 and 2). In the case
of Japan, FDI flows to China in fact have dropped steadily since 1995, both in terms of absolute amount
and the percent of Japan’s total FDI outflow. In the case of the United States, starting from a lower base
than in the case of Japan, FDI flows to China increased from 1995 to 1998 both in terms of absolute
amount and percent of total US FDI outflow. However, 1999 saw a reversal of both trends. Again, it is
simply way too soon to say with any certainty whether these drops signal the beginning of a trend toward
decreased foreign investment from the United States or Japan into China but, nonetheless, the simple fact
that there were declines is not wholly auspicious.
foreign participation. Since 1992, a growing percentage of local affiliates of foreign firms are majority or wholly
owned by the foreign investors, but the term “foreign-invested” continues to apply even to these.
Table 1 Japanese direct investment flow into China, 1996-99
($ US billions or percent)
Flow to China
Percent change from
Flow to all nations
Flow to China as
percent of flow to
Source: Japan External Trade Organization (JETRO).
Table 2 US stock and flow of direct investment to China, 1996-99
($ US billions or percent)
Flow to China
Percent change of
flow from previous
Flow to all nations
Flow to China as
percent of flow to all
Source: US Department of Commerce, Bureau of Economic Analysis (BEA).
3 On this, see the many case examples discussed in Rosen (1999).
The data as indicated in table 2 are, as noted, from Japanese and US sources. In fact, Chinese
sources show significantly larger flows and stocks of FDI into China from the United States than do US
sources and somewhat larger flows and stocks from Japan. To the best of the authors’ knowledge, the
reasons for the discrepancies have not been wholly sorted out but one factor does seem to be that some
US and Japanese firms place their investment on the mainland via subsidiaries in Hong Kong, so that this
investment shows up in the home country data as direct investment in Hong Kong. Thus, table 3 below
indicates flows of Japanese and US direct investment in Hong Kong (1996-99), using home country data.
The US data indicate much volatility, with major increases in flows over the previous year being recorded
in 1996 and 1997, a major drop in 1998, and a recovery in 1999. The Japanese data show, as for the
mainland, a significant drop in FDI flows since 1996-98, but some recovery in 1999.
Table 3 US direct investment flows to Hong Kong, 1996-99
($US billions or percent)
flow to Hong
US FDI flow to
Source: US Department of Commerce, Bureau of Economic Analysis (BEA).
In sum, there is some evidence that the rate of flow of FDI into China during the late 1990s
experienced some slowing down over the levels of the middle 1990s but this evidence is far from
conclusive. What is clear is that FDI flows into China have not continued to grow at the rates that were
evident in recent years. Further, the slowdown (or lack of growth) is unlikely to be the result of saturation
effects. For example, in spite of the large stock of FDI that has entered China to date, on a per capita basis
this stock is in fact not great when compared with other Asian nations, for example, the Chinese stock per
capita is about $160, in contrast to $320 in Thailand and more than $2,000 in Malaysia (Lemoine 2000).
Also, the role of FDI in the Chinese economy by several measures (e.g., value added by foreign invested
enterprises as a percent of manufacturing value added in the domestic economy) is not high relative to
this role in other Asian economies (Lemoine 2000 op. cit.). Further, the sectoral and locational
distribution of this investment is very uneven. For instance, the majority of FDI in China is located in four
coastal provinces (Guangdong, Jiangsu, Fujian, and Shanghai, in descending order) and most of the rest
of this FDI is located in other coastal provinces. Of the residual, the majority is located in provinces
immediately adjacent to the coastal provinces. Thus, vast areas of China, including ones where much
state-owned industry is located, have not been touched by FDI. In spite of the large flows that have come
to China since 1990, there would appear to be ample capacity for China to absorb much more.
Against these figures, it is also noteworthy that FDI approvals by Chinese authorities have risen
in 2000 and 2001 over approvals in 1998 and 1999, suggesting that the trends reported above, if indeed
the data do suggest trends, could be reversed. One interpretation of the rise of approvals is that foreign
investors have been registering intent to invest in China in anticipation of reforms that are likely to
accompany Chinese entry into the World Trade Organization. Whether or not this impending entry will
actually result in realization of the investment is a matter that largely remains to be seen.
Even if investors are becoming discouraged by the policy environment currently prevailing in
China, the emergence of China as a major host nation to FDI has nonetheless been driven by positive
changes in Chinese policy over the last quarter century or so. From 1949 until 1979, China had been
closed entirely to FDI but as part of a series of reforms a partial opening was implemented in 1979 such
that highly limited access to the Chinese market was granted to foreign investors (Rosen 1999 op. cit.). At
that time China sought access to foreign capital and technology but also sought to avoid creating any
competition whatsoever for domestic state-owned enterprises from foreign owned firms. The result was
that only a small amount of FDI entered China over the next 12 years and this was concentrated in sectors
in which either domestic Chinese firms did not participate (hotel development, tourism) or ones where
China urgently needed access to foreign technology (oil field exploration and development). Beginning in
1992, however, further reforms were taken that led to a surge in FDI. Some of the early post-1992 FDI
was concentrated in nontraded goods (e.g., commercial real estate development) but a priority of China
was development of new export industries, and by 1994 about 60 percent of FDI was flowing into
industrial sector activities that were highly export-intensive. As noted above, in more recent years, an
increasing percentage of output of foreign-invested enterprises has served the domestic rather than the
export market but, even so, FDI in China remains concentrated in export-intensive activities. Indeed,
some analysts figure that almost all of the growth of Chinese exports since 1992 can be directly or
indirectly attributed to foreign-invested enterprises (Wei 1995 and 1996, Lemoine 2000, op. cit.).
This is not necessarily however an entirely unalloyed benefit. As developed further in the next
section, foreign-invested enterprises largely engaged in export processing are not well integrated into the
mainstream Chinese economy. Rather, these operations are used to perform labor-intensive operations on
imported goods for re-export. (Or, at least, so goes the usual story. As is developed later, this story itself
might not be wholly correct.)
Even after the 1992 reforms, however, foreign direct in vestment in China remains very controlled
by state policy and, indeed, significant changes in this policy would seem to be indicated by China’s
admission to the WTO. Currently, the central government of China, as well as provincial governments, do
regulate entry of FDI closely or at least attempt to do so. Entry of foreign firms is often conditioned on
the achievement of industrial policy goals as laid out by the state. Foreign firms are most welcome when
these goals cannot be fulfilled by domestic firms. The entry of a foreign firm can be subject to numerous
conditions, for example, such performance requirements as having to use local suppliers, often designated
by the government, or locating in certain areas, or setting up the local operation as a joint venture. Also,
much the same thing can be said about this policy today as was said a few paragraphs back about past
policy, notably that China seeks access to foreign capital and technology but often still seeks to avoid or,
barring complete avoidance, at le ast to regulate competition between domestic enterprises and foreign-
invested ones. Thus, while the aversion to competition has softened in recent years, it has certainly not
entirely gone away.
Arguably, one consequence is that China has to some extent foregone one of the major sources of
benefit of FDI, notably the dynamic gains that come from greater competition. Ironically, to the extent to
which this has happened, China also probably has foregone much, albeit certainly not all, of the benefit
that comes from access to foreign technology. This issue is further discussed in the next section of this
paper. Another possible consequence is even more ironic: by sheltering domestic enterprises, and
particularly the state-owned sector, from foreign competition,4 China almost surely has reinforced the
4 The state-owned sector in China, to be sure, is shrinking in terms of the percent of output originating in this sector.
But, by most accounts, this is only a relative shrinkage; output in this sector is stagnant, but growing in the private
locational advantages with respect to FDI already held by the coastal provinces. This is indeed ironic
because one lamentation in China about foreign investment is that the benefits have largely accrued to the
coastal provinces, with inland areas getting left behind. But the economies of the inland provinces tend to
be more dominated by state-owned enterprises than those of the coastal provinces, and measures taken to
shelter state-owned enterprises from competition by foreigners (and effectively to block foreign takeover
of these) works to the locational disadvantage of the inland provinces. In the final section of this paper,
we argue that the most effective means to draw FDI to inland areas would indeed be to privatize state-
owned enterprises and to allow foreign ownership of these. Again, whether this will happen in the post-
WTO entry era remains to be seen.
The remainder of this paper is divided into three sections. The next section (section 2) examines
what has been the record of FDI to date in China, focusing on a number of recently published studies.
What is revealed, as has already been suggested, is that the benefits have been very great indeed.
However, as also has been suggested, the distribution of the benefits has been very uneven and there
certainly remain potential gains that have not been realized. Section 3 augments these published studies
with some findings of our own, ones that are roughly consistent with the already-published findings.
Section 4 presents some concluding remarks.
EMPIRICAL EVIDENCE OF GAINS TO CHINA FROM FOREIGN DIRECT INVESTMENT:
GROWTH AND EXPORTS
As just indicated, FDI in China has stimulated much growth in income that would almost surely not have
been realized in the absence of this investment, as is demonstrated in two very recent but quite different
working papers (Lemoine 2000, op. cit., and Dayal-Gulati and Husain 2000, henceforth D-G&H). The
former of these is largely descriptive but provides both substantial scope and depth of detail, while the
latter is largely econometric. Together, they demonstrate what by now is a commonly told story about
FDI in China: FDI has significantly benefited the coastal regions while most of the rest of China has
benefited much less. Together, the two studies also provide evidence that this outcome is not wholly the
consequence of natural locational advantages held by the coastal provinces but, rather, is in part the
consequence of Chinese policy.
The core of D-G&H is use of what they call a Mankiw, Romer, and Weil version of the Solow
growth model to test for convergence of income growth to a steady state across provinces of China.5 To
domestically owned and foreign-owned sectors. Many analysts believe that absolute shrinkage of state-owned
enterprises in China must be accomplished.
5 Mankiw, Romer, and Weil (1992). In their empirical tests, D-G&H aggregate the provinces into six regions, north,
northeast, coastal, south, southeast, and west. In this aggregation, the vast majority of FDI is in the coastal region,
with the remainder in the north and northeastern regions.
explore the D-G&H results, including our criticism of these results, some understanding of this model is
necessary. Thus, for the reader who is not familiar with the model, a brief explanation is provided in an
appendix to this chapter.
In fact, D-G&H use a truncated version of the Mankiw, Romer, and Weil version of the Solow
model as described in the appendix. Specifically, D-G&H use the following specification:
ln yit – ln yi0 = C – (1-e- ?t)yi0 + xit + ?it
where the i indexes province, C is “a constant term across all provinces”, xit is a vector of “other
explanatory variables”, and ?it is an error term. As developed in the appendix, in the full Mankiw, Romer,
and Weil model, xit would include terms involving population growth, savings rate, depreciation rate,
human capital investment rate, etc. However, D-G&H omit most of these variables for the simple reason
that necessary data are not available. However, arguably the most important variable, fixed investment to
provincial GDP, is included. But also, D-G&H throw in some other variables that do not figure in the
Mankiw et al. version, notably share of state -owned enterprises in industrial output, share of FDI in
provincial GDP, ratio of public revenue to public expenditure, and ratio of bank loans to deposits. Of
critical importance is inclusion of FDI; the main goal is to test whether FDI has contributed sig nificantly
to the growth of China by province and, in particular, whether the higher per capita growth in the coastal
provinces can correctly be attributed to the high rates of foreign investment there. According to them, the
likely answer is “yes”. As they note, “higher FDI flows could imply more openness, and thereby a higher
rate of convergence, or access to different technology, implying convergence to a higher steady state, at
least in the short run”. However, to test this, they must also look at other explanations for the growth
differentials, thus introduction of other potential explanatory variables.
Their main results are (1) per capita income convergence rates among provinces were not
constant over time (or, more precisely, the hypothesis that these rates were constant over time can be
rejected); (2) the hypothesis of “unconditional convergence” (that per capita incomes were converging to
the same level in all provinces) can be rejected6; and (3) FDI, as expected, both affects significantly and
positively the level of per capita income to which provinces are converging and the rate of income
growth. Regarding the third result, as would be expected if there were unconditional convergence, per
capita income in the poorest provinces of China grew faster than in the richer provinces during the early
1980s, when FDI was not nearly as large a factor in China’s development as it is today. But this was
6 To restate this and hopefully to clarify it, the rates of per capita income growth to which the provinces of China are
converging differ among provinces.
reversed during the 1990s when FDI poured into the coastal areas (as also happened in the northern and
northeastern regions when FDI began to flow there in the late 1990s).
D-G&H also find that the loan-deposit ratio by province has a negative impact on growth,
especially in the time period 1988-97. They interpret this, following Lardy (1998), to be the result of
“policy loans” to state-owned enterprises, that is, state-owned banks in China are forced to lend to loss-
making state-owned firms on noncommercial terms, causing this ratio to rise in those provinces where
such loss-making firms are concentrated. Thus, a high loan-to-deposit ratio would signal that the
government was forcing the intermediation of savings into capital with low returns at the margin in the
affected region. 7 In contrast, they find that ratio of revenue to expenditure by province has little
relationship with per capita income growth. Likewise, the ratio of investment to provincial GDP does not
significantly affect this growth, although inclusion of the variable increases the estimated rate of variance
? as measured by it.
Overall, the D-G&E results show robustly that FDI in the coastal areas of China is positively
associated with relatively higher rates of per capita income growth. They also show, albeit arguably less
robustly, that government policy has suppressed per capita income growth in areas where state-owned
enterprises are concentrated. Somewhat separately from their econometric results, D-G&E also show that
the latter provinces--those where state -owned enterprises are most concentrated–are not the provinces
where FDI is concentrated. A reasonable conclusion would be that for some reason foreign investors shy
away–or are discouraged–from areas where state -owned enterprises are prevalent.8
One can quibble, however, both with the D-G&H model and with the authors’ interpretation of
their results. The most serious flaw of their model is omission of important variables that figure in the
Mankiw, Romer, and Weil version of the Solow model, without which it is not clear that the D-G&H
version is on particularly solid theoretical footing. Importantly, exclusion of these variables prevents them
from performing the sort of “consistency test” that caused Mankiw et al. to reject the original Solow
model as an explanator of differences in real per capita growth rates among nations (this test, as
developed in the appendix, was to determine whether the model yielded a plausible income share of
capital ?). Also, we find ourselves not wholly convinced by the D-G&H’s argument that the relatively
7 Indeed, savings are either used to finance output that is sold at a loss or is retained as inventory; in either case, the
return at the margin is zero or negative.
8 This, however, by itself does not prove that government policies discourage FDI from locating in proximity to
state-owned enterprises either deliberately or even inadvertently. It could be that coastal areas possess locational
advantages for foreign investors but not for state-owned enterprises. Conversely (and as they seem to be true as
artifacts of policy in the early years of the People’s Republic of China), state-owned enterprises might have been
forced to locate operations in areas that were not optimal from a commercial perspective. Chairman Mao, for
example, sought to disperse Chinese industry and to locate it in areas away from the coast for reasons of national
defense and as a way of rewarding those regions of China that had most strongly backed the Communist revolution.