Technology has long been considered as a main driver of economic growth. How knowledge is created has been an emphasis of the new growth theory developed in the beginning of 1990s. Research and development (R&D) refers to an activity that comprises creative work undertaken on a systematic basis in order to increase the stock of knowledge and the use of this stock of knowledge to devise new applications. By investing in R&D, the overall economy can dramatically be benefited in terms of enhance productivity. However, only a few nations around the globe appear to be highly involved in knowledge production. Even if they do, they seem to be allocating lesser resources than what the very huge returns should command.
As a result, two key characteristics of global R&D activity are concentration and underinvestment. Based on the source from UNESCO, less than two thirds of nations around the globe were actively involved in technology production in 2005, and some 25 OECD countries carried more than 86% of total R&D expenditure. During the same period, the empirical evidence on technology development and spillover has witnessed noticeably high rates of return on R&D investment (see: Coe & Helpman, 1995; Jones & Williams, 1998)
If in fact investment in technology enhancement is a key source of economic development over the long term, it is then strange that nations seem to allocate relatively small proportion of their national funds to the R&D sector. This requires a research of the main forces governing the allocation of resources to knowledge production, or the factors that shape the economic and institutional incentives to which inventors and innovators respond. The economic growth literature has introduced some major determinants such as government subsidies (Phelps, 1966; Shell, 1966), knowledge accumulation via learning-by-doing, the size of domestic markets, and trade openness to exploit economies of scales related to technology to fully materialize.
By controlling for the variables of human capital and intellectual property rights (IPRs), this study will be conducted to find out additional possible factors of R&D spending in a nation, namely FDI and financial development in the Malaysian context. The selection of this country is justified by it's comparatively (internationally) low level of R&D expenditure despite the fact that the Malaysian government has allocated quite a huge portion of funds to invest in R&D over the last two decades (MASTIC, 2008).
1.2: Historical Background
Over the past fifty years, Malaysia has achieved noticeable economic transformation via rapid development processes engineered by various economic and national development plans. Since independence, the economy has progress smoothly; starting from a traditional agricultural based economy to a modern industrializing export-oriented nation, and currently, the manufacturing (especially electronics) and the services sectors are playing an important role in taking the challenge of the nation's future economic growth.
In the 1980s, Malaysia began its transformation from a low-income nation into a middle-income nation by gradually diversifying its economic activities. Based on the Tenth Malaysian Plan (2010), its economic growth was stimulated mainly by factor accumulation including capital (investment), energy (cheap fuel) and labor (both local and overseas but primarily low skilled). During the 1980s and 1990s, it transformed from an economy predominantly emphasized on the production of raw materials such as rubber, tin and palm oil into one of the world leading electronics exporters. The sector was the key source of export growth during the transformation period and is now the nation's leading industrial sector in terms of investment, value added, export and employment (MOSTI, 2007).
For a few decades, Malaysia has witnessed rapid economic growth, and before the 1997 Asian financial crisis (AFC), it has been considered as one of the most rapid growing economy in line with other successful East Asian countries. During the period of 1991-1997 (prior to the crisis), Malaysia achieved a real gross domestic product (GDP) growth of more than 9%. During the same period, there was a substantial increase in the per capita income, low inflation rates, and decrease in the incidence of poverty. In particular, in 1990, real per capita income has increased from RM993 to RM4357. It has continuously risen more than twice within 15 years from RM5854 (1990) to RM10004 (2005), and is predicted to ascend further to RM12130 at the end of the ninth Malaysia Plan (MP) period. At the same time, the incidence of poverty reduced from 52.4% in 1970 to 16.5% in 1990 and continuously fell to only 5.7% in 2005 (Yussof, 2009). The rapid growth of the economy resulted in a dramatic structural transformation which led to the gradual switch from great dependence on the production and exports of primary commodities to a more advance industrializing economy.
In short, since independence, Malaysia has achieved tremendous structural transformation that led the country to be able to achieve a remarkable economic development and economic growth. Within this period, agricultural, manufacturing as well as services industries played a pivotal role in taking the challenge of the nation's future economic growth.
Figure 1: Malaysia's GDP Growth during the Period of 1960-2010
-8
-4
0
4
8
12
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
G
D
P
g
r
o
w
t
h
(
a
n
n
u
a
l
%
)
Source: World Bank
In short, figure 1 depicts Malaysia's GDP growth from 1960 to 2010, where the vertical axis represents GDP growth measured in annual percentage term, while the horizontal axis represents years. Generally, there are fluctuations in which the economy grew on one hand, but deteriorate on the other hand. For example, the economy deteriorated during crises such as the 1997-1998 Asian financial crisis (AFC) and the 2007-2008 global financial crisis; while the dramatic increase in growth after both crises may be partly due to the increase in R&D expenditure from 1999 and 2009 as shown in figure 2. However, the growth may not be sustainable and hence may not sufficient enough for the achievement towards the goal of becoming high-income nation. The problem may be related to the moderate R&D expenditure by the government as well as private sectors.
Figure 2: Malaysia's R&D Expenditure during the period of 1980-2010
0
1,000,000
2,000,000
3,000,000
4,000,000
5,000,000
6,000,000
80
82
84
86
88
90
92
94
96
98
00
02
04
06
08
10
R
&
D
E
x
p
e
n
d
i
t
u
r
e
Source: Department of Statistic Malaysia
Empirical literatures have shown that economic growth cannot be separated from R&D expenditure, and that the latter in turn is dependent on FDI and financial development. After the establishment of the seventh five-year plan by the government in 1996, which emphasized the importance of science and technology (S&T) for economic development, Malaysia had witnessed a continuous rising expenditure towards R&D. Figure 2 graphically shows Malaysia's total R&D expenditure from 1980 to 2010, based on the statistics provided by the Department of Statistic (DOS) Malaysia. The vertical axis represents total R&D expenditure (RM in million), while the horizontal axis represents years. Obviously, there is an increasing trend. For instance, there has been a more than five-fold increase in R&D spending during the period of 1980-2010. This shows a steady progress in the nation's R&D activities as a result of the process of globalization. For example, the relatively higher increase in 1999 and 2009 may be due to the concurrently substantial increase in FDI inflows (Figure 3) and domestic credit to private sector (Figure 4) as a percentage of GDP. About one third of the total investments were from the public sector, while the remaining investments were from private sector (Stads, Tawang, and Beintema, 2005).
In particular, based on inputs and outputs of innovation at the national level, as measured by the Ministry of Science, Technology and Innovation (MOSTI), R&D expenditure in Malaysia (as a proxy for inputs innovation) has risen substantially in nominal terms since 1992, specifically in the last decade, from approximately RM0.5 billion in 1992 to more than RM3.5 billion in 2006. But, it has in fact declined in terms of percentage to GDP since 2002, from a peak of 0.7% to GDP in 2002 to 0.6% in 2006. Since 1996, private sector was the main contributor in R&D investment in Malaysia. For example, its R&D expenditure consists of 85% of the nation's total R&D expenditure in 2006 (Tuah, Nadaraja, & Jaafar, 2009).
Malaysia's R&D expenditures are mainly generated from internal sources like governmental budgetary allocations. In 2002, these internal funds constituted a 70% share; foreign sources represented 12% share; 10% came from the Intensification of Research in Priority Areas (IRPA) program; and the remaining was generated from federal and state government funds.
Most of Malaysia's 2002 R&D expenditure was on engineering sciences (39%) and ICT (24%); but agricultural sciences got only 4% (Stads, Tawang, and Beintema, 2005). In the same period, total R&D expenditures constituted 0.69% of Malaysia's GDP, which is higher than some nations in the region (e.g: Thailand at 0.24%; Philippines at 0.11%), but substantially lower than the ratios of India (0.78) and many developed countries (e.g: the U.S at 2.72%; Japan at 3.07%; South Korea at 2.53%) (MASTIC, 2004). Malaysia targets to increase its R&D expenditure to at least 1.5% of GDP by 2010 (MOSTI, 2003).
Despite the fact that Malaysia's R&D spending rose between 1980 and 2010 from about RM1 billion to about RM5.5 billion, it has a long way to go in order to enhance its R&D intensity and to comply with the international standard.
Korea, India, and China which are the world fastest growing countries, are also the most innovative countries. Is the Malaysian R&D expenditure sufficient and efficient? The volume of research indicates the level of R&D expenditure. In Malaysia, the relatively low level of R&D expenditure as percentage of GDP compared to Japan, South Korea, Singapore and Hong Kong can be reflected by the limited number of researcher. There was only a gradual increasing trend in R&D spending over the years. Similarly, there was not much increase in the number of R&D researcher. The ratio of number of researcher to total population in Singapore, however, has risen rapidly over the years. It is harder to increase the number of researcher than to increase the R&D expenditure due to the time and cost in developing research skills and capabilities. Besides, increasing R&D spending followed by the inappropriate number of researcher will lead to ineffective consumption of the spending and low value product development (Tuah et al., 2009).
According to the Malaysian Economic Planning Unit (EPU), although Malaysia targets to raise gross R&D spending (as a proportion of GDP) to 1.5% by 2010, it only spent 0.64% in 2006. But, Singapore spent 2.31% of its GDP on R&D in 2006, and the targeted expenditure is 3% by 2010. Japan (3.32%), South Korea (3.22%), Taiwan (2.58%), and China (1.42%) have relatively higher R&D expenditure of GDP compared to Malaysia in 2006. China (mainland) has increased its R&D expenditure by 50% since the late 1990s and is currently intended to be increased to 2.5% of GDP per annum. Even a sparsely populated country (only 5.2 million citizens), Finland has incurred 3.4% spending of its GDP in R&D, one of the highest percentages in the world (Tuah et al., 2009).
Patenting activity is continuously translated from an economy's capability to innovate. Developed countries have the highest level of patents, but Malaysian firms seldom involve in patenting activity. In fact, Malaysia has relatively low level of patenting activity as compared to Taiwan, South Korea, Hong Kong, and Singapore. In Malaysia, only a small number of patents granted to local firms, while 90% of the patents are granted to foreign firms. Besides, patents granted to local citizens are restricted to semiconductor and electronics industries (Tuah et al., 2009).
Based on the Global Innovation Index 2008-2009, innovation performance (Malaysia) in general and ICT infrastructure are deemed to be unsatisfactory relative to Japan, South Korea, and Singapore. Compared to South Korea, Malaysia has lower internet users and broadband subscribers. Indeed, apart from the declining trend for the period 2000-2006, ICT expenditure as percentage of GDP in Malaysia is relatively lower than Japan, Hong Kong, and Singapore, though it was slightly higher than South Korea. Out of 100 people, Malaysia has 5 broadband internet subscribers, while South Korea 32, Singapore 21, and Hong Kong 28. Since the establishment of the Multimedia Super Corridor (MSC) in 1996, Malaysia is in line with other East Asian countries in broadband penetration. However, a few years after the launch, no substantial improvement exists. Furthermore, the number of broadband subscribers in Malaysia is substantially lower compared to South Korea, Japan, Singapore and Hong Kong where they had a dramatic increase during the period of 1998-2005. While Malaysia shows limited movement. Korea has efficiently utilized broadband penetration to stimulate multimedia-based activities and web services particularly in Seoul (Tuah et al., 2009).
For developing countries, innovation activity is deemed to be an adoption of foreign technologies which needs more adaptation effort to the local aspect. Product innovation means "products new to the firm" instead of "product new to the relevant market". The former category, most common in developing nations like Malaysia, appears as imitative strategy (when taking the world market as a reference point); while the latter means thorough innovation strategy (commonly observed in developed economies). Hence, R&D spending in Malaysia depends on the firm's innovation strategy, that is innovating or non-innovating. This means that firms spend on R&D either to adapt foreign technologies to the domestic market or to improve its competitiveness when facing a stiffer foreign competition without innovating. In this case, R&D activities will not be well organized and could be implicit activity. For example, firms could regard that they don't have R&D activities as they are not innovating; but in fact they are spending on R&D. (Kriaa & Karray, 2010).
As a whole, Malaysia is relatively low in terms of quantity and quality of R&D expenditure, number of R&D researcher, level of patents, and ICT infrastructure as compared to other developed countries such as Japan, South Korea, Singapore, Hong Kong and Taiwan. This reflects its weaker innovative performance than its counterparts.
Figure 3: Malaysia's FDI Inflows during the period of 1980-2010
Source: World Bank
Figure 3 shows Malaysia's FDI net inflows from 1980 to 2010, where the vertical axis represents FDI net inflows measured in percentage of GDP, while the horizontal axis represents time period in year. From the period of 1980-1982, although Malaysia FDI inflows showed an increasing trend (gradual), it was quite inactive due to the lack of knowledge, unpopularity in this area, as well as restrictive government policies. All these will in turn result in less mobility of capital between countries. In 1982 to 1987, there was a decrease of Malaysia FDI inflows before it rose dramatically from approximately 1.2% of GDP in 1987 to 8.8% of GDP in 1992. This was due to the Japanese currency appreciation, Japan and Asian newly industrialized economies' (NIEs) trade friction with the U.S and the EU countries, as well as Japan and NIEs rising wage rates in the mid 1980s (Wong, 2006).
Figure 4: Malaysia's Financial Development during the period of 1980-2010
Figure 4 shows Malaysia's financial development from 1980 to 2010, where the vertical axis represents domestic credit to private sector measured in percentage of GDP, while the horizontal axis represents time period in year. Generally, the trend involves fluctuations from 1980 to 2010. For example, there was a dramatic increasing trend from 1990 to 1997 (correspond to an increasing trend in R&D expenditure during the same period), followed by a substantial decrease from 1998 to 2008 due to the Asian financial crisis (starting from 1997/98) and the global financial crisis (starting from 2007/08). After that, the financial sector bounced back from 2009 to 2010. Again, the rise in the financial sector corresponded to an increase in R&D spending.
In short, it was believed that the relatively substantial increase in R&D expenditure in 1999 and 2009 may be due to the concurrently substantial increase in FDI inflows and domestic credit to private sector (as a percentage of GDP).
1.3: Problem Statement
Sustaining development is likely the most challenging task that countries especially developing ones have ever faced, and achieving it requires the fundamental problems to be solved at domestic, regional as well as international level. At all levels, the role of R&D is important. Innovation and adequate technologies play pivotal role in tackling the economic, social, and environmental issues that lead recent development path unsustainable.
Since independence, Malaysia has experienced rapid growth with an average annual GDP growth of about 6%. The growth was mainly generated from natural resources such as petroleum; and commodities like rubber, palm oil and manufacturing. Land, low-cost labor, and capital served as the traditional driver of growth towards the success.
However, there is a question on whether this growth rate can be increased further or sustainable? In this case, R&D or innovation plays a vital role in moving the economy up the value ladder as it is commonly accepted to be one of the key drivers of economic growth in the knowledge based economy. R&D is important in the sense that it closes technological gap and free developing countries from globalization gap. For the former case, innovation helps in changing comparative advantage and bridging technology gap with industrial economies. While for the latter case, knowledge creation and innovation as a result of R&D activities can move a country from the "user" of capital intensive, high technology, and high value added products to the "creator" of such products.
Although there was a gradual increase in R&D expenditure by both the public and private sectors, the national innovation and creativity level is still insufficient. For instance, under the eighth Malaysian Plan (2001-2005), a large portion of budget has been allocated to raise the knowledge flows. Among them are the expansion of the Multimedia Super Corridor (MSC) flagship project to establish a cyber cities network, the computerization of some ministries and agencies, and the continuing improvement of Malaysia's telecommunications and IT infrastructure. Nonetheless, the main limitation of Malaysia's innovation system is the lack of technically skilled human capital involving in R&D activities, despite the standard indicators of the government's commitment towards human capital efforts, like the portion of total expenditure allocated on education is quite favorable compared to Singapore and even to developed economies like Japan and the USA. In terms of the number of scientist and engineer in R&D and researchers, Malaysia is substantially lagged behind as compared to Singapore, China, and Vietnam. The problem is even worse as a result of the 'brain drain' problem (Chandran, Rasiah, & Wad, 2009). In line with the increase in R&D expenditure, the number of full time R&D staff (a proxy for inputs to innovation) rose during the period of 1992-2006. But, in terms of per 1000 population, it was in fact decreased from 0.7 per 1000 population in 2004 to 0.5 in 2005 (Tuah et al., 2009).
Compared to other nations, Malaysia is leading in terms of the adults and young literacy rate, but that is insufficient to prepare itself towards an innovative economy as both indicators cannot reflect the education quality. The current rapid expansion of new educational institutions also may not guarantee a huge return if the majority of the programs focus on non-science and non-technical subjects as well as if the quality of the graduates decreases. Demand-supply deficit of the high-tech industries is another problem faced by Malaysia. For instance, according to Rasiah (2002), due to ineffective coordination of supply and demand of high end human capital, Penang and the Klang Valley were unable to obtain enough high-tech human capital. In fact, Malaysia's performance is very poor in terms of the tertiary level enrolment ratios compared to other countries, particularly in science and technology. Over the past few years, its international ranking in terms innovativeness and competitiveness has also decreased. Malaysia has known the importance of technology and has started spending a lot in this area. But, without having a higher education system, it is basically impossible to gain (Chandran et al., 2009).
Moreover, Malaysia does not have any specific tools to facilitate or accelerate the positives spillovers generated from the multinational corporations (MNCs) that operate in the manufacturing sector. This is due to the shortage of an advance technology-based small and medium enterprise (SME) sector which can serve as a stable source of supply. This reflects Malaysia's inadequate human resources development policy (Tuah et al., 2009).
The development of new products and processes are the surviving tools of a country, given current competitive environment. Opportunities should be generated from academic research particularly by encouraging partnerships among universities, research institutions, and the private sector. The government should identify and give support to potential research projects in any of these institutions and inside firms in order to facilitate the flow of innovation and new ideas. Patents, licensing, and royalties are the direct outcome of innovation by research institutions and firms efforts. With high patent grants and royalties, Korea and Singapore understand the link between the outcome of the research and economic growth. However, Malaysia is in the low position with respect to this. It is also evident that Malaysia is weak in terms of value added industries (Chandran et al., 2009).
Furthermore, the technological progress of the country is reflected by the firm-level activities in high technology intensives industries. It is also widely accepted that the concentration of linked industries and institutions assist firms in enjoying competitive advantage. The reasons are clustering creates efficient concentration of suppliers, efficient access to information and knowledge, arm-length relationships and coordination, improved diffusion of knowledge on best practices and spurs innovation. But, the pursuit of cluster based industrial development does not lead to a simultaneous growth in the auxiliary industries which will facilitate the engine of growth for the economy. Malaysia is still depending on foreign technology, despite the fact that it has successfully switch to the export of high technology industries. For example, in 2001, there are 57% of Malaysia's export is in high technology industries (Chandran & Veera, 2003; Chandran, Deviga & Karunagaran, 2004). As a result, local technological capabilities have not been substantially stimulated from the spillover effect brought in by the MNCs.
In short, the country's problems in the technological progress measures, coupled with its low level of R&D as well as innovative capability create challenges. Despite the evidenced positive returns from investments in knowledge-based activities, majority firms were still unwilling to involve in these activities. This is due to some limitations like insufficient funds to enhance knowledge capabilities, lack of proficiency in English, insufficient skills, and the uncertainty in the benefit-cost ratio (Tuah et al., 2009). The nation is also suffering from low-skilled jobs and low wages, difficulties in doing business, absence of private investment, and stagnating productivity growth.
1.4: Research Questions
Based on the above issues, the study makes an attempt to answer the following questions: Firstly, does FDI relate to R&D expenditure? If yes, how is it? Secondly, does financial development relate to R&D expenditure? If yes, what is the channel? Thirdly, controlling for the variables of human capital and intellectual property rights (IPRs), are FDI and financial development equally important in explaining R&D spending? Or are they substitutes?
1.5: Objective of the Study
Previous empirical studies on innovation (Hobday, 1996; Rasiah, 2003; Narayanan & Wah, 2000) and internationalization of R&D activities (Ariffin & Figueiredo, 2004) in Malaysia, as well as FDI-growth nexus (Balamurali & Bogahawatte, 2004; Baharumshah & Almasaied, 2009; Batten & Xuan, 2009; Anwar & Nguyen, 2010) witnessed a vast amount of studies over the years. Nevertheless, there were limited studies on the interaction among FDI, financial development, and R&D, although there are deemed to be more growth enhancing. Evidence on this issue is potentially very crucial in understanding one of the key channels through which financial development affects growth. Hence, the general objective of the study is to examine the effect of FDI and financial development on R&D expenditure in Malaysia from 1980 to 2010 by employing the ARDL method.
.
1.6: Significance of the Study
As such, it was hoped that the study can contributes by closing the gap in the existing literature. In particular, it could complements prior theoretical as well as empirical research that have concerned on how FDI interacts with financial development in promoting (relatively higher rate) economic growth. Understanding this element can assist the Malaysian policy makers to solve the existing problems that prevent the economy to move up the economy ladder and hence to achieve the goal of high-income nation.
1.7: Organization of the Study
The next section will be the literature (theoretical as well as empirical) review; followed by section 3, the methodology of the study and data description. Empirical results and interpretation are described in section 4. Finally the conclusion and some implication will be discussed under section 5.
2.0 Literature Review
Prior to the methodology and empirical results, a critical review is a must in order for ones to know what have been done and what have not been done (hence can contribute by closing the gap). In this section, an amount of theoretical as well as empirical review is provided regarding the nexuses of human capital-R&D, intellectual property rights (IPRs)-R&D, as well as the integrated relationship among FDI, financial development and R&D.
While studies have been done on FDI-R&D nexus (see: Feinberg & Majumdar, 2001; Lin & Yeh, 2005; Elmawazini, 2010; Hymer, 1960 (Pioneering work)), and financial development-R&D nexus (see: Hsu, Tian, & Xu, 2010; Nelson, 1959; Schumpeter, 1911 (pioneering work)), there were limited research on the integrated relationship among these variables namely FDI, financial development, and R&D expenditure (see: Maskus, Neumann, & Seidel, 2012). As such, motivation has been given to study on the subject matter.
2.1: Theoretical Review (R&D)
The first literature in economic literature associated with econometric study of R&D activities is Griliches' technical knowledge production function (Griliches, 1979), where it consists of the conventional productive factors and, "technological capital" which depends on firms' R&D spending, universities R&D, and technological centers' activities. This framework has been employed by several scholars (see: Acs, Audretsch & Feldman, 1992; Audretsch, 1998).
However, Griliches' function doesn't take into account all the activities in innovation process that are multidimensional and interactive. R&D is only a component of innovation spending. From the birth of the idea to complete development, and approaching innovation activity by exclusively R&D expenditures involve underestimation, especially in limited-size firms and conventional industries (Calvo, 2000). Indeed, in Calvo's research which took Spanish manufacturing firms as the observation, it was found that most of the process innovation companies did not have R&D staff and more than fifty percent of product innovative firms did not invest actively in R&D.
Accordingly, better models have been introduced to examine the determinants of innovation behavior. One of them is from Camison (1999). He argues that firms' competitive advantage is dependent on resources and capabilities that are hard to copy by other firms. He introduced a model in which organizational characteristics affects innovation behavior, and this in turn influences firms' performance.
After that, a more detailed model of the relationship has been proposed by Hernandez and de la Calle (2006), where firms' size, firms' age, executives' age, enterprising courage, and cooperation with institutions jointly influence innovation behavior, and this in turn will lead to better economic performance, new products and services.
Following Hernandez and de la Calle, Vieites and Calvo (2011) introduced a structural model where different criterions like contingent factors, human, organizational and financial resources, as well as cooperation and information management influence innovation activities. And those activities determine innovation outcomes, influencing firms' performance. The most differencing characteristic of the model is flexibility, breaking the linear structure estimation of the relationship between R&D, innovation and business performance. In this model, there is a more versatile design due to the usage of latent variables (also known as constructs), obtained from observed variables. It also defines various kinds of causality link between those constructs.
On the other hand, a vital component of the growth literature that associated with the endogenous growth theories has studied identical issues within the framework of "ideas" production (Porter & Stern, 2000). One approach regards the number of patents as proxy for knowledge, which is associated with inputs like human capital and the stock of knowledge accumulation. Later studies decompose the latter into local as well as overseas source. Among the main disadvantages of this framework are: by regarding patent citations as a proxy for knowledge, it literally assigns zero knowledge to nations with hidden patents (without declared); these nations might pursue to depend mainly on the spillover or imitation as the key source of technology acquisition. Even for economies with promising and active R&D activities, this framework systematically underestimates the inventory of ideas because several investors might believe that it is better not to disclose their research outcome to the public (non-patenting for instance). As such, they prevent being imitated, and thus induce their competitors to come out with novel, better technologies. Besides, to the extent that it precisely considerate for the stock of patents, this approach will describes more on how productive the R&D sector is and how efficient the operations of the patenting system is, instead of how much of its national resources a nation in fact allocates to the production of ideas.
As a summary, regarding to the determinants of R&D spending, there existed various theoretical models (with their own advantages and disadvantages) that explain the relationship between the former and the latter. However, the frameworks can be modified or improved in order to make more sense and for the explanation of relationship to be more convincing.
This study builds on these frameworks and makes two modifications. Firstly, R&D expenditure is employed rather than patent citations as a proxy for knowledge flow. Secondly, other than financial resources and human (which are associated with financial development and human capital development respectively), additional variables are incorporated to consider for both characteristics (concentration and underinvestment) of international R&D activity. These variables include intellectual property rights and FDI inflows.
2.2: Empirical Review
Although there is limited empirical research studying on the integrated relationship among FDI, financial development and R&D expenditure, there is quite an amount of studies on the nexuses of FDI-R&D and financial development-R&D, in which they can be integrated. There are also literature that concerns on the effect of human capital and intellectual property rights (IPRs) on the R&D activity.
Existing literature indicated that financial market development can solve the problem of asymmetric information and ensure efficient allocation of capital. Therefore, it is crucial to understand whether FDI inflows, complemented by a well-structured financial system, could be an effective source in obtaining greater value of R&D spending.
2.2.1: The relationship among FDI, Financial Development, and R&D
Schumpeter (1942) regarded R&D as a key driver of economic growth and development. He captured the role of innovation in enhancing technology and in contributing towards enhanced efficiency and productivity. Accordingly, a vast number of literatures have existed to empirically test his hypothesis where majority of them has concerned on evaluating the effect of FDI on R&D spending and the role of financial market development in affecting that specific R&D activity. Nevertheless, majority studies of the Schumpeter's hypothesis have been inconclusive and in most cases the scholars were unable to give a convincing explanation that could reinforce or reject the hypothesis. Besides, there are high controversies with respect to the relative importance the innovation and the measure of productivity due to exogenous variables like a nation-wide switch in tax policy, openness of the economy (FDI), skilled researchers and engineers supply (human capital), the institutional framework (financial system) and others, which could influence the R&D decision of a firm. But, there is no common acceptance among scholars regarding which factor to regard and which to exclude. A complete literature review of the empirical research examining firms' R&D decisions can be referred in the work by Cohen and Levin (1989).
The development of technological capabilities is the output of a sophisticated integration of incentive structure with human resource, technology efforts, and institutional factors. The former includes macroeconomics incentives and incentives from competition and factor markets, all of which are associated with the governmental encouragement of FDI inflows; while the institutions consists of market and non-market institutions like financial institutions and legal framework. It is the interaction among all these determinants in specific nation settings that justify, at the regional level, how good the region utilize the resources and develops technological capabilities (Lall, 1992). Besides, it was also found that FDI can contribute substantially as a whole to regional innovation capability, but the strength of this positive effect is dependent on the absorptive capability and the existence of innovation-complementary assets in the recipient region (Fu, 2008). Therefore, other than separately affecting R&D expenditure, FDI and financial development interact with each other in affecting R&D spending. This statement is evidenced by a number of scholars. According to Fu (2008), improvement of domestic absorptive capability has been important to the effective diffusion of the knowledge and technology spillovers from FDI.
As consistently shown in the empirical growth literature, a nation could gain dramatically from its economic partners' efforts to create novel knowledge, for example by hosting foreign technological-advanced activities (Keller, 2004). These spillover benefits (due to labor turnover for instance), known as positive externalities, could give high incentives to establish an R&D sector (Omran & Bolbol 2003; Lin & Yeh, 2005). This is because insufficient competition in a market (due to lacked FDI inflows) will lead to inefficiency and hence a poor innovative activity (Geroski, 1990). In other words, local firms' interactions with more technologically advanced companies (via FDI for example) are more likely to induce them to develop at least equally effective technologies in order to survive and do well in the greater global competition. Cheung and Lin (2004) supported the findings by concluding that there are positive FDI spillover effects on the amount of China's local patent applications, and the effect is the greatest on minor innovation (especially external design patents), either there is no crowding out effect in China or the positive spillover effects dominates the crowding out effects at the provincial level. Given that nations that are in fact at the technology frontier benefit much more from their own R&D investment effort than their counterparts, a nation that is already benefiting from the global technology may intend to enhance those advantages by beginning to invest substantially in knowledge production.
However, if the crowding out hypothesis is being taken into account, instead of innovating on its own, a country can buy technologies from foreigners by establishing joint venture with them (a negative relationship). This substitute is more attractive when the self R&D activities become riskier or costlier and the technology associated is sophisticated (like invention). According to Kathuria and Das (2005), there is a substitutability relationship between FDI and local R&D in the post-1991 era when the effects of FDI are perceived to have been absorbed, though a mix (see: Ciruelos & Wang, 2005; Beladi & Firoozi, 2008; Cheung, 2009) and insignificant (see: Mytelka & Barclay, 2004; Sun & Du, 2010) relationship were also found. By using the model of technology diffusion as a theoretical framework, Ciruelos and Wang (2005) found different effects of FDI on R&D spillovers in developed (positive) and least developed countries (negative); by applying a game-theoretic explanation for the variables' correlation, they stated that both positive and negative correlations between local R&D and FDI inflows can be described in terms of MNCs' competitive and interactive assessment of the efficiency and expenditure consequences of such R&Ds when undertaken by their local competitors; Cheung (2009) indicated that while on one hand, the import of technology is crucial to understand and establish the absorptive capability (and is positively related with innovation performance), on the other hand, crowding out effects do present in a few high- tech industries.
Furthermore, international technology spillover could also demotivate the gaining nations to establish a substantial R&D sector in this way: majority poor nations with relatively tight financial constraints and poor institutions, investing in research (as a result of competition from foreigners) may be not only too expensive, but the returns may be significantly ambiguous (Hyytinen & Toivanen, 2005; Ang, 2010 (by using Schumpeterian growth framework)). Since financing constraints for R&D could be especially tight because of the intangible characteristic of R&D assets. Access to external finance could be predicted to be vital in promoting innovation and hence economic growth because financial constraints retard innovation.
In short, nations' benefits from foreign technology spillover might or might not create enough incentives to establish a promising R&D sector. The former is known as "standing-on-shoulder" effect or international spillover effect which refers to: previous research benefiting existing or future research activity. While the latter is known as the "raising-the-bar" effect, in which foreign discoveries push out the definition of "new" knowledge, as well as the "fishing out effect" which asserts that previous discoveries are easier than the latest ones (Seck, 2011).
However, by decreasing the problem of asymmetric information, a well-developed financial market with the extant legal system can enhance the effectiveness of a nation's R&D activities (as a result of FDI) (Chowdhury & Maung, 2012). Besides, it cannot be expected that financial resource constraints mandatorily cause lower performance of R&D activities. On the other hand, not all innovation activities will fare well when there are financial resource constraints.
2.2.2: The relationship between Human Capital and R&D
On the other hand, while international migration and foreign technical assistance are relatively costless channels through which nations (generally the case for poor nations) can effectively be benefited from novel technologies, the only cost would be to establish high absorption capabilities of foreign technologies. As a result, it would be more attractive to devote (scarce) resources to enhance technology absorptive capability, like investing in higher education (human capital development), or by enhancing the institutional quality (intellectual property rights).
For the former case, economic performance is determined by various macroeconomics policies and structural conditions, and hence is different substantially across nations and regions. Stability-oriented policies (like inflation and fiscal policy), trade policy, financial market conditions and labor market institutions affect the degree to which a growth can be sustainable. Nonetheless, in the long run, economic performance of a nation is also highly determined by knowledge-related variable like human capital. While the literature does not regard innovation explicitly, it could be assumed that highly educated people tend to create more innovations that foster industrial development and hence resulting in economic growth. If one assumes that the level of accumulated knowledge (via spillover for example) positively influence the productivity of local innovative activities, then the greater the knowledge stock, the easier it can produce technology, and the higher the technologically progress in the economy. This was the idea developed by Nelson and Phelps (1966). Hence, they also argue that the optimal capital structure is achievable when the progressiveness of technology and investment in human capital are implementing in the economy. In short, the less educated labor has higher probability to delay or fail to introduce the new technology unlike the more educated labor that can adopt the new technique with success.
There are many reasons to assert that economies with greater absorptive capability are also more probably to establish an extensive R&D sector (positive relationship). Most of the underlying factors that govern resource allocation to R&D on one hand, and foreign technology absorption on the other hand, tend to be similar, for instance highly skilled labor (see: Mishra, 2007; Lee, Florida & Gates, 2010; Kriaa & Karray, 2010; Vieites & Calvo, 2011).In 1961, Jacobs published a book which mentioned that the capacity to innovate is due to a local environment or background that attracts brilliant people that is open and creative. This finding was reinforced by Desrocher (2001) who found that the major obstacle for cities to grow was influenced by the lack of attention and creativity in the cities. Therefore, diversity plays a crucial role in city and regional growth, as creative people from different environment join to generate new and fresh combinations of current technology and knowledge to create innovation. Following the previous researchers, Lee, Florida & Gates (2010) argue that a region's capability to attract human capital and to provide low obstacle of entry for talented and imaginative people of all background determine the capacity to innovate. Hence, human capital and creativity is the major ingredient of innovation.
Besides, recent evidence suggested that the determinants of innovation for Tunisian firms are highly dependent on the human capital quality, past experience in innovation and etc. Besides, the finding also shows that the firm is able to innovate through spillover effect when there is absorptive capacity (Kriaa & Karry, 2010). In line with this view, Cohen and Levinthal (1990) and Daghfous (2004) showed that the absorptive capability is critical for firm to determine the innovative capabilities of advance technology. The absorptive capacity is also known to ascertain the level of research and development. Hence, in order to increase the innovation activity; the firm and organization need to move up their absorptive capacity which can effectively acquire as well as utilized internal knowledge which would affect the firm's ability to innovate. Therefore, with greater absorptive capacity it enables the firm to further raise firm performance of innovation. Diaz-Diaz and Saa-Perez (2012) carried out a similar study to Kriaa and Karry (2010), who examined the relationship of human capital and innovation activity in the Spanish firm level. The authors highlighted that the existing and recently hired R&D human resources is the main determinant to promote innovation activity and performance. Improving labor learning abilities includesthe ability to integrate the peripheral knowledge obtain by employing R&D labor, as well as to improve the in-house knowledge through existing R&D labor to incorporate and develop knowledge from their external environment. This study has undoubted clarify that the human resources is an important foundation of knowledge for product innovation performance. For instance, Vieites and Calvo (2011) found that human and financial resources as well as cooperation influence R&D activities positively. Mazilescu and Mazilescu (2008) supported their finding by indicating that human capital raises the chances of R&D activities. The latter and other Science and Technology (S&T) activities could not be efficiently conducted without human resources. It has conclusively been shown that human capital is one of the important determinants of total factor productivity; it can influence the growth through two channels. Firstly, the capacity of nations to innovate new technologies is directly affected by the level of human capital. Secondly, the higher the human capital accumulation the faster the speed of technological catch-up to the high-tech nation, hence improve the momentum of technology diffusion (Benhabib & Spiegel, 1994).
As a summary, most or even all of the previous empirical literature concluded a significant positive connection between human capital and R&D spending. As such, there is a strong reason to believe that they are positively associated.
2.2.3: The relationship between Intellectual Property Rights and R&D
Intellectual property rights play a key role to protect the creation of the innovator, with the better patent protection to the innovation it give more confident to the investor and innovator hence it is also providing some monopolistic power to the investor. For the case of intellectual property protection, there are a number of studies found that there is to some extent, a positive connection between intellectual property rights (IPRs) and R&D spending (see: Segerstom, 1991; Taylor, 1994; Chen & Puttitanun, 2005; Branstetter et al., 2006; Kanwar, 2006; Fu, 2008; Seifert & Gonenc, 2012).
In 1993, Taylor who examined the North-South model of unintentional technology transfer. He has found that the speeding up of technology transfer would promote competition and raises the productivity of resources employed in the South, unilateral attempts to magnify unintended transfer via weak patent policy call forth defensive responses by innovating firms. Gould and Gruben (1996) highlighted that the intellectual property rights will significantly enhance the activity of innovation through the provision of an environment conducive to the accretion of knowledge. For instance, Fu (2008) indicated that the insufficient of effective IPR protection is a huge obstacle hindering MNCs from introducing greater variety of technology into a country and spending more on innovation. In order to help encourage R&D investment, an effective legal system protection like property protection plays an important role. The author had found that the property rights protection is positive and significantly related to both process and product of R&D (Chen, Ping & Song 2010). Furthermore, the financial deepening and stronger intellectual property protection has an advantage to motive innovation and technological explanation. With effective financial reform in the countries would give stronger confidence to the investors who intend to invest in risky innovation activity (Ang, 2010). The finding was reinforced by Seifert and Gonenc (2012) who indicated that patent rights have a positive impact on R&D intensity.
In contrast, Roffe (1974), Gilbert and Newbery (1982), as well as Zhao (2006) produced results that indicate negative relationship. For example, Zhao (2006) indicates that technologies developed in nations with poor IPR protection are employed more internally, and technologies developed by firms with R&D in poor IPR nations indicates greater internal linkage (negative relationship). The results suggest that firms may use internal organizations to replace inappropriate external institutions. By doing so, they are able to take advantage of the arbitrage opportunities resulted from the institutional gap across nations.
Interestingly, even a mix relationship was found (see: Kim, Lee & Park, 2006; Crampes & Langinier, 2009; Chuang & Lin, 2011).The study of dynamic general equilibrium model of the international product cycle of intellectual property protection (IPP) in South was carried out by Lai (1998), he found that the channel of international production transfer either imitation or multinationalzation (prior to imitation) from North to South is the main factor which affect the effectiveness of IPP. The stronger the IPP, the lower the rate of innovation as the channel of production transfers through imitation. However, IPP has exactly contrary effect when the channel of production transfer is through multinationalization. Besides, when there is a different firm size, Cull and Xu (2005) finds that the property protection rights and access to finance is only effective to smaller firm but not large firm. Thus there is an insignificant relationship between property rights for relatively large firm.
On the other hand, Crampes & Langinier (2009) showed that the set of legal constraints limiting the entry in markets in which an incumbent has innovated should not be similar for all industries. This is because the externalities generated by the entrant are not necessarily negative for the incumbent, or because the incumbent's R&D spending can be detrimental to entrants. Therefore, one particular set of legal rules can on one hand stimulate innovation in some industries, but on the other hand can be detrimental in others. This finding was supported by Chuang and Lin (2011) whose empirical results showed that the IPRs protection in the host nations has a significant impact on the foreign R&D expenditure of MNCs in emerging economies, but it is different significantly between developing and developed nations. A greater extent of IPRs protection in developing economies is found to result in more foreign R&D expenditure; while a stronger extent of IPRs protection in developed nations, especially in industrialized nations, tends to decrease the foreign R&D spending of MNCs in emerging economies.
In sum, there were mixed (positive, negative, and mixed) empirical results regarding the link between IPRs and R&D expenditure. On one hand, IPR would encourage and cultivate competition in the industry which would heighten the productivity of the resources besides the advancement of technology through intensive R&D. On the other hand, there would be relatively less intensive R&D activities should the nation has poor IPR. Further, it can be concluded that there exist an inverse relationship between the IPP and the rate of innovation of a firm. From the above, it can be understood that IPP would not give the same result to the size of firm. The existence of legal constraints in a country would vary with the industries. In other words, the findings are inconclusive. Therefore, it is worthwhile to critically examine the relationship in order to produce a robust result.
CHAPTER 3: METHODOLOGY
3.1: Introduction of methodology
First of all, this chapter shows a theoretical framework (developed based on the theoretical reviews) that provides guidance in explaining the relationship between FDI, financial development and R&D expenditure economically. Besides, the model specification and econometric technique employed in the study are introduced, followed by data description and the source in the next subsection. The nature of unit root tests is also briefly explained before the depiction of the relevant diagnostic checking in the last.
Specifically, in terms of econometric techniques, the concept of cointegration initiated by Granger (1981) as well as Granger and Weiss (1983) before it was extended and modified by Engle and Granger in 1987 is incorporated. The existence of a stationary and long-run equilibrium relationship among two or more time series is explained by the notion of cointegration, though they are individually non-stationary (Narayan & Narayan, 2004). One of the main benefits provided by cointegration method is that it facilitates us in integrating the short-run and long-run relationships between two or among more than two variables in a framework that is unified. Furthermore, the spurious regression problem can be partially or totally solved by the existence of cointegration among the time series. Under the notion of cointegration, the autoregressive distributed lag (ARDL) approach which includes the bounds testing procedure was employed.
Before proceeding to cointegration test, unit root tests such as Augmented Dickey Fuller (ADF) test, and Philips-Perron (PP) test were conducted to examine the data property (stationarity).
3.1: Theoretical Framework
In this subsection, rather than looking at the relationship between FDI and R&D expenditure as well as the FD-R&D nexus individually, FDI, FD and R&D are integrated in order to provide a guidance or direction to analyze their inter-relationship in the following section. This is because the extent of the FDI impact on innovation capacity may be relied on the absorptive capacity (namely financial soundness) of the host countries. Note also that nations that spent more on R&D are more likely to innovate because R&D directly creates new products and production processes (Fu, 2008). As a result, the words 'R&D' and 'innovation' are used interchangeably.
Based on various literatures, there is a strong relationship between domestic financial market development and R&D intensity. According to Omran and Bolbol (2003), technological innovation and resource allocation can be improved via better savings mobilization and the direct effect of savings on capital accumulation. However, among sources from the international capital markets, only FDI seems to be a key factor in funding R&D (Maskus, Neumann, & Seidel, 2012). FDI generates competition and linkages for indigenous producers. This induces them to improve their technologies that need funds, especially for new entrepreneurs who face financial constraints. In this case, a developed financial system that enables access to external source of fund as well as better distribution and surveillance of these funds is crucial. It is also vital in justifying the degree of MNCs' borrowing to expand their innovative activities into the local economy. Moreover, the evolution of domestic financial sector, which serves as a platform to channel funds from surplus units to shortage units, and diffusing technology embodied in FDI inflows play a pivotal role in affecting the speed of technological innovation as well as pattern of economic growth of a country.
In contrast, an underdeveloped financial system may hinder the economy from gaining benefits from the potential spillover effects of FDI. For example, if a domestic entrepreneur wants to integrate and practice the best technology generated by FDI, a poor financial sector restrict the potential positive FDI externalities (Alfaro, Areendam, Sebnem & Selin, 2006).
In summary, FDI and financial development may separately as well as jointly affects R&D expenditure, though the latter's effect is perceived to be more substantial. This is because the extent of the FDI effect on innovation capacity may rely on the absorptive capacity (namely financial soundness) of the host countries. Therefore, it is hypothesized that FDI and financial development positively affect R&D expenditure.
3.2: Model Specification and ARDL approach
To examine the impact of FDI and financial development on R&D spending in Malaysia, the econometric method employed in testing cointegration property among the selected time series is the ARDL procedure advocated by Pesaran and Shin (1995) (see also Pesaran & Pesaran 1997; Pesaran, Shin, & Smith 2001) and it is includes the bounds testing approach. In addition, the notion of cointegration was first established by Granger (1981) as well as Granger and Weiss (1983).
The bound testing approach has the underlying computed F-statistic (Wald F-statistic), that is used to examine the significance of lagged levels of the variables under the study in a conditional unrestricted equilibrium correction model (UECM). This approach has a few advantages as compared to others like Johansen and Juselius (JJ) tests and Engle-Granger two-step method. One of them is: regardless of the stationarity of the independent variables (whether they are in I(0), or I(1), or mutually cointegrated), it is applicable. As a result, due to its independency on pretesting the variables' order of integration, it avoids the risks associated with the pre-testing of order of integration. Further, the ARDL can avoid finite sample size problem which is faced by JJ tests and Engle-Granger two-step method. In other words, it is more applicable in research that involves small sample size (this study consists of only 31 observations) as compared to JJ tests and Engle-Granger two-step procedure. The ARDL can also test both long run and short run relationships of the estimated variables.
The Vector Auto-Regression (VAR) of the order p (VAR(p)) for Malaysia R&D expenditure function must be demonstrated when the ARDL procedure is employed:
Where is defined as the vector of both and, where is the endogenous variable (measured by Malaysia R&D expenditure) and (foreign direct investment, financial development, human capital development and intellectual property rights) is the vector matrix of a set of exogenous variables. 't' is measure the time or trend variable. The in the model must be I(1) variable, but the can be either I(0) or I(1) according to Pesaran et al. (2001). VECM (Vector Error Correction Method) can be further developed as below:
Where Δ=1-L. Choong et al (2005) developed the long run multiplier matrix and are now separated as:
λ=
The crossway elements of the matrix are unrestricted. Thus, the series can be chosen either I (0) or I (1). If , then y is consider as I(0). In contrast, if , the y series can be considered as I (1).
The ARDL method involves two stages (Narayan & Narayan, 2004). Firstly, the existence of a long run relationship is established. Secondly, a two-step procedure is employed to estimate the long-run relationship. To investigate whether a long-run relationship is present in equation or not, the unrestricted error correction (UEC) models have to be estimated. Specifically, to examine the impact of FDI and financial development on Malaysia's R&D expenditure, the following unrestricted error correction model (UECM) of the ARDL model will be estimated:
The existence of long-run relationship in the model is tested by using F-test. According to Narayan & Narayan (2004), the F-test will indicate whether which variable should be normalized when there is long-run relationship in the model. Besides, the null hypotheses and alternative hypotheses for the model are constructed as following:
(There do not exists long-run levels relationship)
(There exists long-run levels relationship)
The hypotheses above can also be indicated as (Y|X)
The bound test approach that is used under this research follows the F-test non-standard distribution which depends on (a) whether the UEC model is with drift and / or a trend or not (b) the number of exogenous variables, and (c) whether the variables included in the UEC model are one order of integration I(1) or zero order of integration I(0) (Narayan & Narayan, 2004). Since there are only 31 observations (less than 100), the critical values based on Narayan and Narayan (2004) are used. In order to select an adequate lag into the model, Schwartz Bayesian Criteria (SBC) was being referred.
To examine the impact of FDI and FD on Malaysia's R&D expenditure in a more specific manner, the following model is the unrestricted error correction model (UECM) of the ARDL model:
where the ε1t in the equation is the disturbance for the ARDL model. Besides, DUM represents time dummy which capture the Asian financial crisis in 1998. The null hypothesis that testing the long run relationship of the empirical model for the first equation is and for the second equation. While the alternative hypothesis contradicts with the null hypothesis as it states that at least one βj or (j=1, 2, 3, 4) not equal to zero, it means at least one variable has long run relationship.
If the computed F-statistic (Wald test) of bound test is higher than the upper bound critical value, we can reject the null hypothesis and conclude that the variables are cointegrated. However, if the F-statistic is lower than the lower bound critical value, we cannot reject the null hypothesis and we can conclude that there is no cointegration in the model. Another possible outcome is that, if the value of F-statistic is reclining between the lower bound and upper bound value, we can conclude that the cointegration property is indeterminate (Narayan & Narayan, 2004). Under the ARDL approach, Schwarz Bayesian Criteria (SBC) was employed to select the optimal lag length of the model.
3.3: Date description and data source
Variable
Definition
