The Effects of Inward FDI on Domestic Innovation

(1)

The Effects of Inward FDI on Domestic Innovation

- Evidence from China

Thesis: MSc International Economics and Business

Author: Li, Hao

Student number: s1941194

E-mail address: (h.li.5@student.rug.nl)

(2)

Abstract

In recent decades, technological progress driven by innovative activity is

considered as a fundamental ingredient which promoting sustained

economic development. In this paper, I attempt to find whether or not

inward FDI leads to such technological progress- innovations. By

adapting the empirical framework of national innovative capacity and

employing panel data analysis under fixed effect model, I find that inward

FDI itself does not facilitate domestic innovations in China. The main

drivers of innovation in China are local human capital and domestic

knowledge stocks. Besides, local human capital and domestic knowledge

stocks not only directly contribution to innovations, but also enhancing

domestic absorptive capacity. The positive spillover effects of FDI could

be achieved if sufficient absorptive capacity exists.

(3)

‘Economic growth’ is not only an issue in reality- developed countries devote efforts to sustain their prosperity and less developed countries attempt to catch up with their counterparts, but also an attractive issue in academic research. As Grossman and Helpman (1994) pointed out, if we deciphered what drives technological progress, then we can pursue a faster and sustained economic growth. Coincide with this point of view, many scholars’ interests shift towards to the determinants of technological progress. Grossman and Helpman (1991) stated that technological progress is comprised of two components: innovations and imitations. Innovation refers to ‘creating new ideas’ while imitation refers to ‘ideas obtained by means of knowledge

dissemination’. Since technological progress driven by innovative activity is widely

recognized as a fundamental ingredient that promotes long- term economic development, what we need to know is how to promote innovations.

(5)

engage in innovative activity and promote their innovative performance. Nevertheless, this point of view considers firms as a main actor in innovative activity and ignores other potential actors which engage in innovative activities. As a complementary research, Nelson (1993) devoted efforts to explain innovation in a way of national innovation system. Most of literatures based on this argument are descriptive studies or case studies. This point of view was inspired by the fact that innovation process depends on the interactions between all active agents within innovation system. Innovation is results from collaborations and competitions among all these actors. Furthermore, the approach of innovation system emphasizes the importance of institutional factors which stimulate innovations.

Despite from different angles, contributions given by all above are mainly focusing on the internal factors (factors pertain to a nation) which have impacts on innovations. In this paper, I intend to find out the whether external factor that has influences on innovations- inward Foreign Direct Investment (FDI). To my knowledge, researches investigated relationship between inward FDI and innovations are remaining relatively scarce. To fill the gap, the research question of this paper is:

Does inward FDI have impacts on innovations of China? Are there any factors interacted with FDI and jointly impacting innovations of China

(6)

and Wang (2003) find that the presence of multinational firms leads to the increase of total factor productive (TFP) in Chinese industries. Borensztein, Gregorio and Lee (1998) and Mello (1999) provide similar optimistic findings by investigating cross- country sample. Nevertheless, the increased productivity or technological progress may stems from innovative activity or imitative behavior (creating new technology or adopting advanced technology). Moreover, the increased TFP can driven by using exist technology more efficiency. In this sense, although they convinced that host countries technological progress benefits from FDI, they failed to prove that inward FDI lead to genuine innovations in host countries. Therefore, there are merits to examine whether host countries’ role really changed from imitators to innovators (Hu and Mathews 2008). In other words, the nexus between inward FDI and innovations of host countries are incompletely explored. Secondly, from the perspectives of China: amongst developing countries, China has become one of the largest importers, exporters and FDI destinations. In fact, FDI inflows to China escalated from around $100 million in 1979 to over $60 billion in 2004 (Hale and Long 2007). Besides, multinational firms set up over 700 R&D centers in 2004, the number of MNEs R&D centers increased to 1100 by the end of 2009. These facts suggest that multinational firms have decentralized their R&D activities to China. Bruche (2011) stated that China has become one of most attractive destinations of MNEs’ R&D centers. Qun and Peng (2006) argued that technology diffusions in China are mainly caused by inward FDI. Moreover, in recent decades, China attempts to establish an innovation-oriented society (Suttmeier, Cao and Simon 2006). These facts also make China to become a more suitable country to investigate whether inward FDI promotes domestic innovations.

(7)

paper attempts to find whether FDI leads to ‘pure’ technological progress- innovations.. 3. The absorptive capacity of China is taken into consideration in the innovation context. To answer my research questions, I employ panel data analysis. The dataset covers 28 provinces over 11 years. The empirical analysis of this paper is based on the national innovative capacity framework and incorporating the factor of inward FDI.

This paper is organized as following: In section 2, I review relevant literatures and based on which the hypotheses are formulated. Section 3: methodology and data. Section 4: empirical results. Section 5: Discuss the results. The last part is conclusions and limitations.

2. Literature review and Hypotheses:

In this section, I review two strands of literatures. One group of literatures focused mainly on internal determinants of innovations, the other group of literatures focused on the effects of inward FDI on host countries. Then I establish linkages between these two strands of paper and formulate hypotheses.

2.1 Literature review:

2.1.1 Theories of Innovation:

(8)

idea generation activities, more productions of new ideas. The other factor is the availability of existing ideas to the manpower. Existing knowledge which is available to these human capitals guides them to finding new idea and increases their productivity. Moreover, this model assumes knowledge spillovers. In other words, manpower in ‘idea sector’ has equal opportunities to access to the knowledge pool which was generated in previous periods of time. Although Furman, Porter and Stern (2002) argued that this specification is highly abstractive, it offers insights to fundamental drivers of innovation.

(9)

customers’ requirement. Moreover, a firm has close bond with other firms (intra- and inter- industries) may facilitate knowledge spillovers. With these new knowledge or skills, firms are able to engage in an array of innovative activities. Finally, since local firms compete in domestic context, competition conditions in domestic market create pressures for local firms. Fierce competition leads firms to reduce their cost or to upgrade their technology by innovation in order to survive. Based on this point of view, we can understand why some firms are more competitive or innovative in a nation or region.

Nevertheless, firm is one of those actors but not the only one which engages in innovative activities. This leads us to the approach of systems of national innovation to understand innovative activities. National innovation system theory argued that innovation process results from and depends on the interactions between all active agents within national innovation system (Nelson 1993). Generally, the national innovation system consist of institutions which in charge of formulating technology and innovation policy, organizations which performs and finances R&D activities, organizations and polices which promote human resource augmentation, and interactions between above factors (OECD 1999). Innovation is a function of all actors who potentially contribute to innovation and promote knowledge flows. Mytelka and Barclay (2004) argued that the rationale of ‘national systems of innovation’ approach is to understand knowledge flows between different firms, organizations or institutions. Besides, they stated that knowledge flows not only occurred along the value chains, but also other actors outside value chains that cannot always be detected in advance. Chang and Shih (2004) also stated that innovation process is not isolated activity. Therefore, it is desirable to examine all active actors surrounded.

(10)

perspectives into consideration. The contributions of this framework is not to provide new theories, it transforms above three strands of theories into a testable framework. By adapting this framework, a series of empirical had been done (OECD: Furman et al 2002, East Asia: Hu and Mathews 2005, China: Hu and Mathews 2008).

2.1.2 Theories of FDI:

Theories of innovation mainly explained internal factors and mechanisms that affecting innovations in a given country. However, innovations and local conditions may also influenced by external factors. In this section, I elaborate how inward FDI affects domestic innovation.

Multinational firms enter foreign markets to exploit their firm-specific advantages as they are recognized as most R&D-intensive actors who possess advanced technologies and knowledge (Markuson 2002). These superior technologies and knowledge are related to new product, producing technologies and management skills (Blomstrom and Kokko 1998). When multinational firms mere presence in host countries, they may unintentionally transfer their superior technologies and knowledge to host countries. This is also the reason why many countries compete to attract foreign direct investment. Regarding how the technology and know-how is transferred and why FDI is crucial for innovative activities in recipient countries or firms, the mechanisms are derived from FDI- Economic growth theory.

(11)

(12)

profits they potentially obtained. Last but not least, the benefits of FDI are not only captured by firms, but also the embraced by entire region (Fu 2008). Innovation is a process of decision making, funding, managing, designing, producing and marketing. Multinational firms own rich experiences on entire innovation process. Therefore, when local organizations or institutions collaborated with multinational firms, these experiences are learnt by actors within national innovation system.

2.2 Hypotheses:

2.2.1 FDI and Innovation:

All above mechanisms guide us to propose that FDI may contribute to innovations. Besides, we also need empirical evidence to support theoretical predictions. Not all empirical researches are done in macro level, however, empirical evidences in firm level are also helpful to understand and support theories listed in previous section. Based on theories and together with empirical findings, I list my hypotheses.

(13)

to produce existing goods in a more efficient way) and product innovation (firms are able to launch new products which lead to increase in varieties). They find that FDI not only allows transferring of new product which more likely affect product-innovation, but also may lead to process-innovation through transferring technology and know-how. Moreover, the author also suggested that the indigenous firms turn out to be more innovative after foreign penetration can be ascribed to intensified competition in domestic market.

(14)

which is opposite to codified knowledge. Moreover, tacit knowledge is vital in interpreting codified knowledge. Therefore, knowledge spillover is geographically confined due to tacit knowledge is hardly transferred to a long distance. In other words, geographical proximity matters with knowledge dissemination and assimilation (or spillover). Leamer and Storper (2001), they argued that innovative activities rely on transfer of complex tacit message. Geographic distance raises the costs of transferring knowledge and technology. Storper and Venables (2004) argued that knowledge spillovers can be utilized more efficiently in lower geographic distance. Besides, face to face contact still play a vital role in exchanging information that may contribute to innovation. Greenaway (2002) further suggested that the vertical technology spillovers more likely occurred if suppliers and multinational firms are in the same location. Based on these arguments, several seminal literatures contribute to this perspective. Fu (2008) stated that FDI has significant spillover effect on regional innovation capabilities in developing country. Besides, the extents to which FDI may generate such effects depend on the absorptive capacity and the availability of innovation-complementary assets in host region. Cheung and Lin (2002), by using provincial data in China, they found that inward FDI has significantly positive spillover effect on domestic innovation and highlighted the ‘demonstration effect’ of FDI on innovation. Most of these articles convince the positive effects of FDI on innovations.

Based both on these theories and the empirical findings:

H1: Keep other factors constant, inwards FDI has positive effect on innovations in

China

2.2.2 Human capital, Knowledge stocks and Innovation:

(15)

determined by local endowment. In Kinoshita (2001), the author proposed a concept of two faces of R&D. It implies that R&D investments affect firms’ productivity, not only by promoting innovation and technological upgrading, but also enhancing firms’ ability to assimilate and adopt external knowledge. By the same token, I consider human capital and knowledge stocks share the same characters with R&D investments and absorptive capacity is more important in context of innovation.

In Romer’ specification (1990), human capital in idea sector directly relate to technological progress. To coincide with Porter’s framework, human capital can be considered as advanced factors which result from investment in human being. Both arguments deem manpower as a driver of innovations since innovation is recognized as knowledge intensive activities. Dakhli and Clercq (2004) classified human capital into firm specific, industry specific and individual specific human capitals. All types of human capital are those who possess certain skills or expertise obtained from educations or job training. With certain amounts of ability embed into them, they could engage in more complex activities. In a country level or regional level study, individual specific human capital is more suitable since it carries a broaden sense than firm specific and industry specific human capital. Barro and Lee (2000) argued that human capital measured by educational attainment has direct effect on economic progress through the skills they possessed. Although they did not directly link human capital to innovations, the argument also indicates the vital role of human capital in technological progress. The research of Teixeira and Fortuna (2003) focused on Portugal confirmed that human capital contribute to increases of productivity. They also state that human capital plays a more significant role when producing process becomes more knowledge intensive. To the case of China, by taking advantages of substantial growth rate disparities among provinces, Fleisher, Li and Zhao (2010) found that human capital had significant effect on TFP growth. They interpreted human capital as an engine of generating innovation.

(16)

channel to transfer technology, by using human capital accumulation to measure absorptive capacity, they find that the host countries benefit from spillovers of FDI only when the human capital exceed certain threshold. Girma (2002) used threshold regression techniques to confirm that the absorptive capacity is crucial. Without sufficient absorptive capacity, there is no positive spillover effects exist. Lai, Peng and Qun (2006) found that the magnitude of technology spillovers from FDI relies on China’s human capital. Liu and Buck (2007) investigated in high-tech industries in China. They suggested that the interaction between indigenous firms’ absorptive capacity and technology spillover determines the innovation performance of domestic firms. Therefore, summarizing their contributions, human capital is also vital in capturing the spillover effect of FDI. With sufficient human capital, it is more efficient to absorb and utilize advanced technology and transform into ‘new to the country’ or even ‘new to the world’ knowledge.

H2: Keep other factors constant, human capital has positive effect on innovations

in China

H3: Keep other factors constant, the magnitude of the spillover effect of inward

FDI on domestic innovation depends on the human capital level. Higher level of

human capital enhances the positive effect of FDI on innovation.

(17)

not well supported empirically. Another point of view suggested that technological progress and productivity of research and development moves alongside with knowledge augmentation. In other words, previously discovered knowledge guides people to generate new knowledge (Porter and Stern 2000). Therefore, knowledge stocks not only offering the amount of existing knowledge that can be exploited, but also enhancing the ability to explore new knowledge (Hu and Mathews 2005). Empirically, by using GDP per capita and patent stocks as proxies, Furman et al (2001), Hu and Mathews (2005) found that knowledge stocks lead to innovation in OECD and East Asia, respectively. Hu and Mathews (2008) failed to find the relationship between knowledge stocks and innovations in China by using GDP per capita as proxy, but the relationship exists after correcting measurement issue (using patent stocks alternatively).

(18)

capacity in capturing spillover effect of FDI.

H4: Keep other factors constant, knowledge stocks have positive effect on

innovations in China

H5: Keep other factors constant, the magnitude of the spillover effect of inward

FDI on domestic innovation depends on the knowledge stocks level. Higher level of

knowledge stocks enhances the positive effects of FDI on innovation.

3. Methodology and Data:

3.1 Empirical approach:

The original model used in empirical research was provided by Furman et al (2002). The model includes a bunch of factors which determine the innovative capacity in country level. By definition, ‘National innovative capacity is defined as country’s

potential – as both an economic and political entity – to produce a stream of commercially relevant innovations’. A group of empirical study was conducted based

on this framework (OECD: Furman et al 2002, East Asia: Hu and Mathews 2005, China: Hu and Mathews 2008, Eastern Europe: Krammer 2009). There are three categories of factors playing roles in determining the innovation performance in country level: Common innovation infrastructure (resource commitments, policies, human capital and cumulative knowledge), cluster-specific environment for innovation, and the linkage of above two jointly affect innovative capacity in a country (the role played by universities). In essence, this framework treats each country as distinct system of innovation.

(19)

according to following reasons: All provinces are under the control of central government, but each province is able to autonomize their local innovation plans. Moreover, labor mobility between provinces is subject to resident permit and each province is characterized differently in terms of historical, cultural and geographical dimensions, which help to shape the innovation system in each region. Therefore, each province formed distinct regional innovation system. In this sense, the national innovative capacity framework can be applied in provincial level in China.

Based on above, the general statistical model used in this paper is formed as:

Log(Domestic Innovation)i,t+1 =

βi +βt +β1Log(FDI)i,t +β2Log(human capital)i,t +β3Log(knowledge stock)i,t

+β4Log(FDI)i,t*Log(human capital)i,t. +β5Log(FDI)i,t*Log(knowledge stock)i,t +

β6controlsi,t + εi,t

All variables are transformed into logarithm form except those measured in percentage, which is consistent with knowledge production function form and less sensitive to outliers (Furman et al 2001, Hu and Mathews 2005, Krammer 2009). In appendix 1, we see that after transform into logarithm form, the data are concentrated. Besides, since no variable takes non-positive value, there is no observation loss after transforming into logarithm. The subscript i denotes province and subscript t denotes year. β1, β2 and β3 capture the effect of FDI, human capital and knowledge stock on domestic innovation. β1, β2 and β3 are all expected to be positive, which is correspond to hypothesis 1, 2 and 4. β4 and β5 express that the magnitude of effects of FDI on innovation depends on human capital level and knowledge stock level respectively. β4 and β5 are both expected to be positive, which is consistent with hypothesis 3 and 5.

(20)

instrumental variables. With strong instruments, we can obtain the variations in independent variable which are determined exogenously. Extant literatures do not provide a valid instrumental variable. However, the bias of this type endogeneity problem can be mitigated by lagging independent variable of FDI as used in previous research (Fu 2008, Krammer 2009). The innovation performance in year t+1 will not affect the FDI inflow in year t. This rule can also be applied to other independent variables.

3.2 Data source and measurement:

The data is collected from: online database of ‘National Bureau of Statistics of China’; online database of ‘State Intellectual Property Office of People’s Republic of China’; China Science and Technology Statistical yearbooks. In this research, all data is provincial- level data which covers 11 years. There are 31 provinces in China nowadays. In this paper, Qinghai, Xinjiang and Tibet are excluded due to data limitations. Therefore, the sample includes 28 provinces of China as listed blow in alphabetic order:

Anhui, Beijing, Chongqing, Fujian, Gansu, Guangdong, Guangxi, Guizhou,

Hainan, Henan, Heilongjiang, Hubei, Hunan, Inner Mongolia, Hebei, Jiangsu,

Jiangxi, Jilin, Liaoning, Ningxia, Qinghai, Sichuan, Shandong, Shanghai, Shanxi,

Shaanxi, Tianjin, Yunnan, Zhejiang.

Measuring output:

(21)

obtain such tax credits, enterprises misreport information on their new products on purpose (Cheung and Lin 2004). In this paper, I select granted domestic patents as a measurement of domestic innovation. We have to bear in mind that patents is not perfectly measuring innovations since patents are different in terms of their economic importance and value (Hu and Mathews 2005). However, patent is one of the most useful measurements of innovations in extant research. Furthermore, I consider that using patent to measure innovations of China has advantages. Firstly, the patent data of China includes both process and product innovations which outperformed than new product sales (Cheung and Lin, 2004). Secondly, each province complies with uniform patent laws in China. Therefore, it is consistently measuring innovation across provinces. Finally, Fu (2008) stated that granted patents in China are categorized by their novelties. There are three types of patents: Invention, Utility model and External design. Invention is characterized as ‘novelty, inventiveness, and

practical applicability, which is a new technical solution relating to a product, process, or improvement thereof;’ Utility model, ‘which means a new technical solution relating to the shape or structure of a product that is not directly related to its aesthetic properties;’ External design, ‘which involves a new design of shape, pattern, or combination, or of color or aesthetic properties’. The distinctions allow us

to investigate the effect of inward FDI on different types of patents. Based on above definition, we perceived that only invention and utility model patents related to new technology and facilitate technological progress, while external design patent focused only on superficial novelty (Li 2009). In this sense, two types of most technological sophisticated innovations are treated as dependent variables:

1. Granted domestic ‘invention’ patent (ip). 2. Granted domestic ‘utility model’ patent (ump). They are measured in ‘unit’ from 1999 to 2009

Measuring input:

(22)

1997. This measurement is consistent with previous research (Cheung and Lin, 2004).

Knowledge stocks (ks): self-calculated by using perpetual inventory method based on summation of invention and utility patent data of each province. This method had been widely used in previous research (Megna and Klock 1993, Krammer 2009). The procedure is as following: Let S refers to knowledge stocks. P denotes patent flows. δ denotes depreciation rate of knowledge. Previous researches indicate that the result is not sensitive to the magnitude of depreciation rate. In this paper, I take the value of 15% which is normally used. The knowledge stock in province i and year t is computed as Si.t= (1-δ) Si, t-1 + Pi.t. In this equation, we need to obtain the knowledge

stocks of the base year, which is calculated as S0= P0/ (g+ δ0). Let δ0 refers to the

average growth rate of patent in first ten years. Historically, Hainan province was established in 1988 and the first patent was granted in 1989. Therefore, the base year is 1989 due to patent is normally granted one year later in China. δ0 is average growth

rate from 1989 to 1998. To launch the calculation, I also conducted somewhat data-correction work. Chongqing was part of Sichuan province until the year end of 1996, so the patent data of Sichuan (from 1989 to 1996) is corrected as Patent data of entire Sichuan minus that data of Chongqing. Knowledge stocks data is calculated for each province from 1998 to 2008.

(23)

equivalent scientist and engineers as proxy ignores the person who engaged in informal R&D and rules out the possibility of that people come up with new ideas suddenly, but it is consistent with Romer’s specification and recognized as the most effective measurement in the context of innovation (Romer 1990, Furman et al 2002, Mathews 2005). In this paper, I use Full Time Equivalent Personnel for Research and development Activities in each province to measure human capital level. FTE personnel for R&D include the full time equivalent scientist &engineers and other R&D related workers. Moreover, this measurement reflects human capital ‘stocks’ but not ‘flows’, which render it to be more suitable. The data of FTE personnel for R&D in each province spans from 1998 to 2008.

Controls:

(24)

3.3 Methodology:

Based on above measurements, two statistical models are formed as (Capital L denotes logarithm):

Lipi,t+1 =

βi +βt +β1Lfdii,t +β2Lftei,t +β3Lksi,t +β4Lfdii,t*Lftei,t. +β5Lfdii,t*Lksi,t + β6Lpopi,t

+β7UNIrd+β8Lrd+εi,t (1)

Lumpi,t+1 =

βi +βt +β1Lfdii,t +β2Lftei,t +β3Lksi,t +β4Lfdii,t*Lftei,t. +β5Lfdii,t*Lksi,t + β6Lpopi,t

+β7UNIrd+β8Lrd+εi,t (2)

(25)

any regressors, which is relaxed in fixed effect estimations. To statistically choose between fixed effect model and random effect model, Hausman test provides powerful information. However, the unit of the observations is each province which cannot be treated as random. Moreover, no time invariant variable within each panel is needed to be estimated and fixed effect model allows the individual heterogeneity correlated with explanatory variables. Therefore fixed effect model are selected.

Before launching empirical test, the quality of data should be examined. To conduct a regression analysis, the assumptions on regression model must be hold. Therefore, some routine diagnostic check needed to be performed prior to regression analysis. Firstly, it is important to check whether errors or dependent variable is normally distributed. In this paper, dependent variables are transformed into logarithm. Therefore, the distribution of dependent variable can be considered as with normal distribution. The test results are listed in appendix 2.

Secondly, multicollinearity occurred when variables strongly correlated with each other or variables lack of variation. If the variables are perfectly correlated with each other, which implies one or more independent variables are exact linear functions of other independent variables (in correlation matrix, it is equal to 1), the assumption of regression model is violated. Moreover, if independent variables are highly correlated with each other, it is difficult to estimate their effects on dependent variable separately. The correlation matrix is listed below:

UNIrd -0.0458 0.0358 0.0112 0.0187 -0.0232 1.0000 lpop 0.3368 0.5582 0.5503 0.4203 1.0000 lrd 0.7498 0.9361 0.9309 1.0000 lks 0.7657 0.9481 1.0000 lfte 0.6807 1.0000 lfdi 1.0000 lfdi lfte lks lrd lpop UNIrd

(26)

and knowledge stocks (lks), 0.936 and 0.931 respectively. Besides, knowledge stocks (lks) are strongly correlated with FDI (lfdi) and almost perfect correlated FTE R&D workers (lfte). In this paper, I propose that FDI and knowledge stocks have influences on innovation. Furthermore, innovation is measured by patents and knowledge stock is calculated based on patent data. Therefore, the correlation between FDI and knowledge stock is expected. Besides, human capitals are crucial factors that facilitate to generate new knowledge. Therefore, it is not surprising to obtain such a high correlation between FTE R&D workers and knowledge stocks either. There is no effective methodology to reduce the multicollinearity. To deal with it, I have to drop one or more variables when run regression analysis.

Moreover, Breusch- Pagan test and White test which are both based on Lagrange Multipliers can be used to detect heteroskedasticity issue. Heteroskedasticity violates the underlying assumption that each random error has the probability distribution with the same variance. In this paper, since fixed effect model have been selected, Modified Wald test would be more suitable, which used to detect groupwise heteroskedasticity in fixed effect model. Autocorrelation of panel data is checked by using Wooldridge-test (2002). This approach is able to test the first order autocorrelation in panel data by using xtserial syntax. The advantage of this test is that it is able to detect the serial correlation only for idiosyncratic error (Wooldridge 2002). Based on the results in appendix 4, the autocorrelation has been detected. To make the regression result to be convincing, I use heteroskedasticity and autocorrelation consistent errors to account for above problem.

4 Empirical Results:

(27)

them can be dropped. To deal with this problem, each of them enters the regression alternatively. In table 1, the dependent variable is granted domestic invention patent which is considered as most technological sophisticated innovation in China. In table 2, the dependent variable is granted domestic utility model patent. In model (a) and model (c), I include FTE R&D workers (human capital) and the interaction between FTE R&D workers and FDI as regressors. While the knowledge stocks and the interaction between knowledge stocks and FDI are used in model (b) and model (d). The estimation is based on entity and year fixed effect model (year dummies are included). Model X1 is the result without interaction term while the interaction term enters model X2 (X: a b c d). The output tables also report the number of observations and R-squares. The R-squares in each regression is sufficient high. It indicates that the data fits well and a large variation in dependent variables is explained. In the following paragraphs, I will mainly report the results which associate with my research interests.

Invention patent as dependent variable:

(28)

and hypothesis 4 is supported. In model (b2), when interaction term between FDI and knowledge stock is added, the coefficient of the interaction term is significant with a positive sign. It indicates that knowledge stocks also increase the absorptive capacity as the results reported. In summary, the positive effect of inwards FDI on domestic innovation is not found. While hypotheses 2 to 5 are confirmed.

Utility model patent as dependent variable:

(29)

Table 1: Dependent variable – granted domestic invention patent

model(a1) model(a2) model(b1) model(b2)

FDI -0.023 (0.070) -1.262 (0.289)*** 0.045 (0.041) -0.425 (0.157)** FTE R&D workers 0.509 (0.197)** -0.544 (0.282)* Population 1.521 (1.189) 1.032 (1.076) 0.325 (0.464) 0.049 (0.391) University performed R&D -0.341 (0.314) -0.126 (0.199) -0.299 (0.152)* -0.194 (0.142) FDI*FTE R&D workers 0.129 (0.029)*** Knowledge stocks 1.501 (0.140)*** 0.908 (0.226)*** FDI*knowledg e stocks 0.056 (0.019)*** _cons -12.979 (9.564) 1.055 (8.229) -11.060 (3.560)*** -3.946 (4.295) R2 0.91 0.93 0.95 0.95 N 308 308 308 308 a: * p<0.1; ** p<0.05; *** p<0.01

(30)

Table 2: Dependent variable- granted domestic utility model patent

model(c1) model(c2) model(d1) model(d2)

FDI -0.007 (0.050) -0.633 (0.234)** 0.037 (0.033) -0.113 (0.125)

FTE R&D workers 0.354 (0.102)*** -0.178 (0.190) Population 0.462 (0.773) 0.215 (0.730) -0.254 (0.345) -0.342 (0.308)

University performed R&D -0.259 (0.256) -0.150 (0.192) -0.242 (0.106)** -0.208 (0.098)**

FDI*FTE R&D workers 0.065

(0.023)*** Knowledge stocks 0.949 (0.083)*** 0.760 (0.141)*** FDI*knowledge stocks 0.018 (0.013) _cons -0.090 (6.186) 7.001 (5.873) 1.100 (2.644) 3.368 (2.897) R2 0.89 0.90 0.94 0.94 N 308 308 308 308 a:* p<0.1; ** p<0.05; *** p<0.01

(31)

5 Discussions:

Based on above regression results, I can conclude that FDI has no independent positive effect on domestic innovations in China, which is opposite to hypotheses 1 and previous research such as Cheung and Lin (2002) and Fu (2008). This could be interpreted as the coefficient of FDI measuring the net effect (the difference between positive and negative effect). The positive effect of FDI on innovations stems both from spillover effect or competition effect, while the negative effect of FDI on innovations is caused by competition effect. When technology gap between foreign firms and majority of local firms is large, increased competition intensity may lead local firms (except those with higher productivity) less likely engage in innovative activities (Aghion et al 2005). Moreover, since local firms may subject to lower efficiency in R&D process, allocating more ‘energy’ to innovative activities will not make them successfully compete with foreign firms. Therefore, if FDI cannot be used effectively, the external superior knowledge cannot smoothly flow within the entire systems. Without efficient knowledge flows, the entire innovation system failed to benefit from it. However, this phenomenon can be reversed only when the region possess sufficient absorptive capacity, which is indicated by interaction terms in regressions. In this paper, the absorptive capacities are related to abundance of human capital and knowledge stocks. They are important to directly generate new ideas and absorb external knowledge. The former is consistent with Romer’s (1990) specifications, and the latter increases the importance of them in innovation context.

(32)

interpreted as the inherent linkage between universities and industries is weak. Universities probably can not provide industries what they need. This may also suggests that the public R&D and industry R&D could be more effective in facilitating innovations in China.

6 Conclusions and Limitations

(33)

(34)

References:

Aghion, P., Bloom, N., Blundell, R., Griffith, R and Howitt, P. 2005. Competition and innovation: An inverted U relationship. Quarterly Journal of Economics, 120: 701–728.

Albert, G. Z., Hu. and Jefferson. G. H. 2002. FDI Impact and Spillover: Evidence from China’s Electronic and Textile Industries. The World Economy

Alfaro, L. 2003. Foreign direct investment and growth: Does the sector matter? Harvard Business School

Ashghian, P. 2004. Determinants of economic growth in the United States: The role of foreign direct investment, The International Trade Journal, 18:1, 63-83

Barro, R.J., Wha, L.J., 1996. International measures of schooling years and schooling quality. American Economic Review 86:2, 218–223.

Bernard, A. B., Jensen, J. B., Redding, S. J and Schott, P. K. 2007. Firms in International Trade, Journal of Economics Perspectives, 21: 105-130.

Bertschek, I. 1995. Product and Process Innovation as a Response to Increasing Imports and Foreign Direct Investment, Journal of Industrial Economics, 43: 341-357.

Blomstrom, M. and Kokko, A. 1998. Multinational corporations and spillovers, Journal of Economic Surveys, 12: 3, 247–277

(35)

Bruche. G. 2011. A new geography of innovation- China and India rising, FDI Perspectives

Carkovic, M. and Levine, R. 2005. Does Foreign Direct Investment Accelerate Economic growth? University of Minnesota, Working Paper

Chang, P. L. and Shih, H. Y. 2004. The innovation systems of Taiwan and China: a comparative analysis. Technovation, 24: 529-539

Cheung, K. Y. and Lin, P. 2004. Spillover effects of FDI on innovation in China: Evidence from the provincial data. China Economic Review, 15: 25-44.

Cohen, W. M, and Levinthal, D. A. 1990. Absorptive capacity: a new perspective on learning and innovation. Administrative Science Quarterly, 35:128–152.

Cornes, R. and Sandler, T. 1986. The Theory of Externalities, Public Goods, and Club Goods, Cambridge: Cambridge Univ. Press

Crespo, N. and Fontoura, M. P. 2006. Determinant Factors of FDI Spillovers- What Do We Really Know? World Development, 35: 3, 410-425

Dakhli, M. and Clercq, D. 2004. Human capital, social capital, and innovation: a multi- cross country study, Entrepreneurship and Regional Development: An International Journal, 16: 2, 107- 128

(36)

Fu, X. L. 2008. Foreign Direct Investment, Absorptive Capacity and Regional Innovation Capabilities: Evidence from China. Oxford Development Studies, 36: 89-110.

Furman, J. L., Porter, M. E. and Stern, S. 2002. The determinants of national innovative capacity, Research Policy, 31: 899–933.

Furman, J. L. and Hayes, R. 2004. Catching up or standing still? National innovative productivity among “Follower” Nations, 1978–1999, Research Policy, 33: 1329–1354.

Girma. S. 2003. Absorptive capacity and productivity spillovers from FDI: a threshold regression analysis. European Economy Group Working Papers

Girma, S., Gong, Y. D. and G örg, H. 2008. What determines innovation activity in Chinese State-owned Enterprises? The role of foreign direct investment. World Development, 37: 886-873.

Gorodnichenko, Y., Svejnar, J. and Terrell. K. 2008. Globalization and Innovation in Emerging Markets, IZA Discussion Paper, 3299

Grant, R. M. 1991. Porter’s ‘competitive advantage of nations’: An assessment. Strategic Management Journal, 12: 535-548.

Greenaway, S., Upward, R. and Wright, P. 2002. Sectoral and geographic mobility of labour markets and structural adjustment, Mimeo, University of Nottingham

(37)

Grossman, G. M. and Helpman, E. 1994. Endogenous innovation in the theory of growth, The Journal of Economic Perspectives, 8: 23-44

Hale, G. and Long, C. 2007. Is there evidence of FDI spillover on Chinese firms, productivity and innovation? San Francisco. CA. Federal Reserve Bank of San Francisco, Working Paper, No. 2006-13.

Howells, J. R. L. 2002. Tacit Knowledge, Innovation and Economic Geography, Urban Studies

Hu, M. C. and Mathews, J. A. 2005. National innovative capacity in East Asia, Research Policy, 34: 1322–1349.

Hu, M. C. and Mathews, J. A. 2008. China’s national innovative capacity, Research Policy, 37: 1465–1479.

Javorcik, B. S. 2004. Does foreign direct investment increase the productivity of domestic firms? In search of spillovers through backward linkages. The American Economic Review, 94: 605–627.

Katila, R. and Ahuja, G. 2002. Something old, something new: A longitudinal study of search behavior and new product introduction, The Academy of Management Journal, 45: 6, 1183- 1194

Keller, W. 2009 International trade, foreign direct investment, and technology spillovers, Handbook of the Economics of Innovation, Amsterdam: Elsevier-NorthHolland, pp. 793-829

(38)

Kokko, A., Tansini, R. and Zejan, M. C. 1996 Local technological capability and productivity spillovers from FDI in the Uruguayan manufacturing sector, The Journal of Development Studies, 32: 602.

Krammer, S.M.S. 2009 Drivers of national innovation in transition: Evidence from a panel of Eastern European countries. Research Policy, 38: 845-860.

Krammer, S.M.S. 2010 International R&D spillovers in emerging markets: the impact of trade and foreign direct investment. The Journal of International Trade and Economic Development, 19: 591.

Lai, M. Y., Peng, S. J. and Qun. B. 2006 Technology spillovers, absorptive capacity and economic growth, China Economic Review, 17: 300- 320.

Leamer, E. E. and Storper, M. 2001. The economic geography of the internet age. Working Paper 8450, National Bureau of Economics Research, Washington.

Li, X. B. China’s innovation capacity in transition: An empirical approach, Research Policy, 38:338-357

Liu, X. H. and Wang, C. 2003. Does foreign direct investment facilitate technological progress? Evidence from Chinese industries, Research Policy, 32: 945- 953

Liu, X. H. and Buck. T. 2007. Innovation performance and channels for international technology spillovers: Evidence from Chinese high-tech industries. Research Policy, 36: 355-366.

(39)

Mello, LR. Jr. 1999. Foreign direct investment-led growth: evidence from time series and panel data, Oxford Economic Papers, 51: 133- 151

Mytelka, L. K. and Barclay. L. A. 2004. Using Foreign Investment Strategically for Innovation, The European Journal of Development.

Nelson, R. 1993. National Innovation Systems: A Comparative Analysis. Oxford University Press, New York, Oxford.

OECD. 1999. Managing National Innovation Systems. OECD, Paris

Porter, M. 1990. The Competitive Advantage of Nations. Free Press, New York.

Porter, M. E. and Stern, S. 2000. Measuring the ‘Ideas’ Production Function: Evidence from International Patent Output,” mimeo, MIT Sloan School of Management.

Qun, B. and Peng, S. J. 2006. A Comparative Panel Data Estimation on Technology Spillover Effects of FDI and Import in China, International Conference on Management Science and Engineering

Rodriguez, C. A. 1996. Multinationals, linkages, and economic development. American Economic Review, 86: 4, 852–873

Romer, P. M. 1990. Endogenous technological change, The Journal of Political Economy, 98: 71- 102

(40)

Storper. M. and Venables. A. J. 2004. Buzz: face to face contact and the urban economy. Journal of Economic Geography, 4: 351

Suttmeier, R. P., Cao, C. and Simon, D. F. 2006. Knowledge innovation and the Chinese Academy of Sciences. Science China, 312: 58

Wooldridge, J.M. 2002. Econometric Analysis of Cross Section and Panel Data. MIT Press

(41)

Appendices:

Appendix 1a: summary statistics: original data

trd 308 91115.86 135298.8 816.2195 849958.1 unird 308 8477.399 11050.95 12.92567 81686.5 fteper 308 42251.33 38894.63 848 238683.8 pop 308 4460.451 2516.376 538 9717 ks 308 11553.8 12741.15 220.1306 93250.23 fdi 308 2520.319 3780.019 16.8 25120 ump 308 3136.451 4300.839 46 27438 ip 308 688.8799 1294.393 6 11355 Variable Obs Mean Std. Dev. Min Max

Appendix 1b: summary statistics: in logarithm (except Unird which measured in %)

UNIrd 308 .1069606 .0758012 .0118993 .9962134 lrd 308 10.543 1.440369 6.704683 13.65294 lpop 308 8.196049 .7151315 6.287858 9.181632 lump 308 7.397678 1.196617 3.828641 10.21968 lks 308 8.838971 1.096024 5.394221 11.44304 lfte 308 10.21492 1.067447 6.742881 12.38289 lfdi 308 6.775743 1.630298 2.821379 10.13142 lip 308 5.615958 1.341928 1.791759 9.337414 Variable Obs Mean Std. Dev. Min Max

According to about result, we can observe the variation of variables. Therefore, all explanatory variables can be estimated under fixed effect model.

Appendix 2: Normality of dependent variable:

lump 308 0.0612 0.4171 4.18 0.1235 lip 308 0.3637 0.8299 0.88 0.6448 Variable Obs Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2 joint Skewness/Kurtosis tests for Normality

(42)

Appendix 3: correlation matrix –include dependent variables UNIrd 1.0000 UNIrd UNIrd 0.0167 -0.0135 -0.0458 0.0358 0.0112 0.0187 -0.0232 lpop 0.3442 0.5585 0.3368 0.5582 0.5503 0.4203 1.0000 lrd 0.9385 0.9104 0.7498 0.9361 0.9309 1.0000 lks 0.9041 0.9752 0.7657 0.9481 1.0000 lfte 0.8544 0.9183 0.6807 1.0000 lfdi 0.6873 0.8126 1.0000 lump 0.8735 1.0000 lip 1.0000 lip lump lfdi lfte lks lrd lpop

Appendix 4a: Autocorrelation in panel data: invention patent as depend variable

Prob > F = 0.0000 F( 1, 27) = 37.297 H0: no first-order autocorrelation

Wooldridge test for autocorrelation in panel data

Appendix 4b: Autocorrelation in panel data: utility model patent as depend variable

Prob > F = 0.0000 F( 1, 27) = 99.866 H0: no first-order autocorrelation

Wooldridge test for autocorrelation in panel data

(43)

Appendix 5: Map

Note: one of those red dots which located in the west of China indicates the location of Sichuan and Chongqing. Chongqing province was established in 1997. Hainan province which located in the South of China was established in 1988. This information is used to calculate patent stocks. This map has nothing to do with political issue. Unit in this research (provinces):

Anhui, Beijing, Chongqing, Fujian, Gansu, Guangdong, Guangxi, Guizhou,

Hainan, Henan, Heilongjiang, Hubei, Hunan, Inner Mongolia, Hebei, Jiangsu,

Jiangxi, Jilin, Liaoning, Ningxia, Qinghai, Sichuan, Shandong, Shanghai, Shanxi,

The Effects of Inward FDI on Domestic Innovation