High Technology Trade and the Roles of Absorptive Capabilities in Malaysia

Abstract

This article analyses the importance of high-technology trade as a channel of economic growth to ease Malaysia out of the middle-income trap. This study also wonders upon the missing absorptive capabilities that validate the likelihoods of dismal gross domestic product (GDP) growth since the 1990s. Using the autoregressive distributed lag (ARDL) approach and fully modified ordinary least squares (FMOLS) estimator as robustness checks, this study identifies the determinants of high-technology trade and the appropriate absorptive capability in enhancing economic growth. The empirical results from quarterly data from 1990 to 2015 proved that foreign direct investment, financial development and infrastructure are vital to develop a successful high-technology trade. Another important finding is that it validates the presence of trade openness (as absorptive capability) in order to magnify the benefits of research and development (R&D). This explains why, despite spending on R&D, these spending do not project to economic growth.

Keywords

High-technology trade absorptive capacity FDI R&D trade openness

I. Introduction

In recent decades, the high-technology manufacturing and high-technology trade have been the fastest growing areas of world trade, which are likely to have a significant impact on the overall economic development of the nations. The term ‘high technology’ is widely used to refer to any firm or industry that embodies products or services with the most innovative and advanced technologies (Seyoum, 2004). Such firms often display a common reliance on sophisticated scientific and technological expertise and rely heavily on research and development (R&D) expenditure relative to turnover (Keeble & Wilkinson, 2000). High-technology trade involves exports and imports of products under the Standard International Trade Classification (SITC—Rev. 1),¹ and the Organization for Economic Co-operation and Development (OECD) defined it as the manufacture of technical products with high R&D intensity. High-technology exports of Malaysia have shown a progressive pattern and are expanding every year in terms of export value but have an inconsistent trend in terms of its contribution to total export. It has successfully contributed to an average of half to the nation’s total export for the past three decades (1996–2016). According to European Statistical Office (Eurostat), Malaysia is among the top three countries in the world, which has the highest percentage of high-technology trade to total trade.

High-technology trade has been the lifeblood of a nation. Towards the transformation into a high-income country, Malaysia has banked on innovative science and technology initiatives that are built on the new economic model (NEM), government transformation programme (GTP) and economic transformation programme (ETP) since 2010. The government defines high-income threshold at a gross national income (GNI) per capita of about US$15,000, which follows the definition by the World Bank. The International Monetary Fund (IMF) and the World Bank have repeatedly called for structural reform and endogenous innovation to move the country up the value chain of manufacturing, hence, allowing Malaysia to leap out from the current middle-income trap. The middle-income trap is generally associated with the notion that countries are stuck in a certain range of income distribution and could not reach high-income status (Cherif & Hasanov, 2015). Based on the World Development Indicators (WDI), Malaysia’s GNI per capita has a steady growth pattern. The GNI per capita is recorded at US$6,530 purchasing power parity (PPP) in 1990, US$11,880 PPP in 2000, US$20,020 PPP in 2010 and US$28,681 PPP in 2017. Malaysia’s GNI per capita has steadily increased throughout the years, but the performance of the economic growth is quite dismal. Although the Malaysian government put in an enormous effort into economic planning to increase the economic growth, it only managed to sustain the growth on an average of 5% since the 1990s.

Malaysia’s economy is in need of a breakthrough in the income per capita and economic growth to achieve high-income nation status. For this, the Malaysian government has accentuated on the role that high-technology trade could play in enhancing the economic growth. The expansion of high-technology trade offers the country opportunities to improve productivity and enhances job creation to further increase income per capita of the citizens (Roukanas & Karakostas, 2019). Therefore, first, this study identifies the potential factors or the determinants of high-technology trade. Upon examination of the direct effect from the determinants of high-technology trade, the study evaluates the indirect effect of high-technology trade on economic growth. As the projection in gross domestic product (GDP) growth since the 1990s does not exhibit a breakthrough, the study touches upon the validation of missing absorptive capabilities that enhance high-impact growth to the economy. Given the facts that (a) the nature of high-technology industry embodies high R&D intensity and (b) Malaysian economy still depends heavily on foreign technology through foreign direct investment (FDI), the study investigates the interaction between R&D and FDI with the presence of absorptive capabilities. Traditionally, human capital is used as the proxy for absorptive capabilities. This study adds two new proxies for absorptive capabilities, which are trade openness and economic freedom, to investigate the indirect effect of the Malaysian high-technology trade through the interaction with R&D and FDI.

All of these efforts into innovation, constant learning and related trade policies would not be successful without the presence of a nation’s absorptive capability. The term ‘absorptive capability’ by Abramovitz (1986), which involves ‘…various efforts and capabilities that developing countries have to develop in order to catch-up, such as improving education, infrastructures and, more importantly, technological capabilities’ (Fagerberg & Godinho, 2005, p. 523). Studies on the national system of innovation have focused on the capability of the economy to adopt and develop new technologies (Haddad & Harrison, 1993; Harbi et al., 2009; Kim, 1980; Mowery & Oxley, 1995). The successful development of high-technology sectors plays an important role in the creation of national welfare. Successful examples of early starters as high-income countries, such as Hong Kong, Singapore and South Korea, serve as an inspiration to carefully study the importance of venturing into high-technology industries. Observing the disappointing performance of GNI per capita and GDP growth of Malaysia, this study suggests that the role of absorptive capabilities could be the reason why economic growth could not reach a higher level. Three absorptive capabilities, namely human capital, trade openness and economic freedom, are used to interact with R&D and FDI to examine the missing indirect effect that boosts the performance of high-technology trade.

Hence, the study contributes to the literature in these aspects. First, this study validates the role of high-technology trade to economic growth and defines the determinants of high-technology trade, specifically, in the case of a developing country, by using Malaysia as a benchmark since its GDP growth is dismal and falls under the middle-income trap. As past literature have focused on developed countries (Falk, 2007; Falkowski, 2018; Fotros & Ahmad, 2017), it is not wise to compare their success with developing countries, whereby technology and knowledge assimilation is still mild. Second, upon deciding on its direct effect by identifying the determinants, the study moves into indirect effect to incorporate missing absorptive capabilities as the reason as to why GDP performance of Malaysia could not leap out of the middle-income trap. Third, the study tends to validate the importance of not only developing the main factors of high-technology trade but also to validate the ‘factors-creation’ that enhance economic growth. Optimistically, policy implications could be to channelize correctly and bring Malaysia out of the middle-income trap.

The rest of this article is organized as follows: Section II reviews the related literature, theoretical framework and hypotheses development. Section III explains the empirical model, the data and econometric methodology, and discusses the empirical results and interprets the findings. Section IV concludes the article.

II. Literature Review and Research Methodology

Empirical Literature

Innovation-based activities and business based on economic motives, generally, are the main factors for technological advancement and economic growth (Pajooyan & Faghihnasiri, 2009). The rationale of the endogenous growth model is based on the form of R&D, in which emphasis is laid on endogenous technology for long-term growth. Technological innovation in human resources and R&D efforts reflects on the storage of knowledge, which, in turn, is used in the production of final goods that result in the increment of production growth rate (Grossman & Helpman, 1991).

By and large, the expenditure on R&D to increase the diversity and quality of products has sparked the interest of this study and stimulates growth in two ways: (a) direct way through innovation and (b) indirect way through enhancing the ability of absorption and transfer of technology. When a recipient country buys an intermediate or final product, the recipient country also buys the R&D that is hidden inside the product (Falk, 2007). The interaction between foreign businesses or FDIs increases internal production as it exposes domestic production with foreign science and technology that is not available, and it is a non-cost-paying recipient of information (Fotros & Ahmad, 2017). International trade patterns in high-technology markets are determined by comparative advantages. However, these comparative advantages would be affected by resources that different economies can devote to industrial R&D (Falkowski, 2018), specifically, to investment in the creation of new knowledge (Rong et al., 2020). Therefore, the relationship among high-technology trade (HTRD), innovation (R&D) and FDI is highly emphasized, as suggested by this study on Malaysia. Although not each and every piece of new industrial knowledge affects international trade, whether it does or not depends on the subsequent diffusion of this knowledge.²

According to Mani (2000), R&D in developing countries should be accompanied, first, by the intensification of the globalization process. Given the revolution in telecommunications, nowadays, the manifestation into global trade is definitely much easier. The second driver occurs with rapid technological change. Mani (2000) demonstrated that the old industrial era is fading out to give way to a new profound development phase. This new era will present developing countries with both opportunities and challenges—challenges that require more knowledge and entrepreneurial spirit to accentuate the development process under a strong global competition. Globalization has, indeed, changed the nature of competitive advantage. Production process has been de-centred; global production networks are connected with interactive developments from the information and communications technology (ICT) industries (Porter, 1985). Globalization presents more locational choices—multinational companies (MNCs) find it cost-efficient to locate the stages of production in different host countries.

One important impact from globalization, not to be missed, is the direct investment in a developing country from MNCs. The original model of FDI and technological impact involves two steps (Martin & Bell, 2006). Step one starts with the international transfer of technology from the parent MNC to the subsidiary, presuming that the parent MNC introduces a superior body of technology than the prevailing technology in the host economy. Step two involves a technological spillover effect, which introduces the subsequent spread of the superior technology to domestic firms. In conjunction with that, a number of studies (Batten & Vo, 2009; Grossman & Helpman, 1991; Guo et al., 2020; Hermes & Lensink, 2003; Teixeira & Fortuna, 2010) have suggested the vital contribution of FDI in a modernizing economy and promoting growth to the host nations. Given the fact that large R&D shares are undertaken by MNCs, FDI from these firms could undeniably become a potential channel of access to advanced technologies available on the global marketplace (Guimon et al., 2018; Kayalvizhi & Thenmozhi, 2018; Zhu & Jeon, 2007). Chuang and Lin (2010) find that Taiwanese MNCs that engage in higher levels of domestic R&D activities engage in overseas R&D activities. Taiwanese MNCs that engage in overseas R&D activities in less developed countries raise their domestic R&D activities.

While these theoretical models of FDI and technology transfer are well built, empirical studies still yield varied results (some with positive results—Zhu & Jeon, 2007; some with no evidence—Xu & Wang, 2000). Several others argued on the institutional circumstances of the host countries (Hermes & Lensink, 2003) to explain the growth effects of FDI. FDI then positively reacts to economic growth only when the economy has achieved a certain level of absorptive capabilities—AC (Alfaro et al., 2004; Borensztein et al., 1998; Sultana & Turkina, 2020). Root (1992) once said multinationals, indeed, transfer more technology, and even more R&D, via FDI to developing economies as compared to technical assistance and help programmes initiated by local governments and international agencies. Along with sufficient technical support, MNCs enable the linkages of local science institutions with foreign affiliates (domestic firms with international networks) through subcontracting and other arrangements (Cantwell, 1999; Patel & Pavitt, 1998). Azman-Saini et al. (2018) demonstrate that firms in developing countries are more inclined towards imitation of the existing products rather than innovation of a new technology. Nevertheless, domestic R&D activity appears to benefit from imports of machinery and equipment,

The Schumpertian terms could be used to support the need to develop the high-technology sector (Harbi et al., 2009; Kroll & Liefner, 2008; Schumpeter, 1934). According to his argument, innovation conducted by high-technology-intensive industries foster entrepreneurial rents. In other words, a country has an advantage of earning higher returns on investments. This includes the benefit from exporting technology-intensive products, which enjoy more stable prices, higher rents and profit margins, and positive and dynamic externalities (Archibugi & Pietrobelli, 2003). Hence, establishing a high-technology sector is a striking proposition to developing countries. Interestingly, there are some successful cases among developing countries, especially the Asian countries. This triumph seems to be challenged by numerical misapprehension that these countries are specializing in the labour-intensive elements of high-technology production. As argued by Kumar and Siddharthan (1997), the bulk of R&D in developing countries is adaptive rather than creative and, therefore, does not carry the same weight into sustainable competitive advantages. That could be one of the reasons why high-technology sectors in developing countries have been overlooked.

This study hypothesizes the role of high-technology export in projecting high-quality growth that is well needed for emerging countries. Numerous studies have shown the possibility of catching up with projected growth using high-technology manufacturing such as the once carried out by Seyoum (2004), Fotros and Ahmad (2017), Harbi et al. (2009), Ismail (2013) and Harris and Li (2009). If growing exports can lift a nation’s economic prospects, then expanding high-technology exports is necessary as it is associated with high value-added production, success in foreign markets and high compensation levels. Seyoum (2004) emphasizes on high-technology industries to create a large share of innovation (including new products or processes that help accumulate market share), create new markets entirely or may lead to more productive use of resources when ‘factor-creating mechanisms’ (AC) are present.

Innovation includes a non-linear learning process taken when diverse actors and networks interact at different levels. Joseph (2013) celebrates economic growth with the effect of trade liberalization. However, it is only helpful with the presence of ‘inclusive growth’—technology, infrastructural investment, education, R&D, etc. Hence, high-technology export exposes local actors to foreign actors and network, which comprise superior technology. With the presence of AC, learning by doing cultivates innovation activities by fostering higher entrepreneurial rents and competitiveness in international market. Eventually, it supports the growth of the local economy. Carolan et al. (2013) use the same SITC trade data from this study and show that the country with the most advanced technology trade basket exhibits the largest trade growth. Japan was dominant in the study as a result of its ability to ‘invent and innovate’ and eventually moving towards capital-intensive goods.

Multinationals also stress on the presence of infrastructure (INF) before they make investments (Athreye & Cantwell, 2007; Tang & Zhang, 2016). Therefore, competence-creating subsidiaries have to be closely surrounded by local networks (Birkinshaw et al., 1998) to facilitate the exchange of knowledge between local institutions in these networks to create a virtuous cycle of innovation growth, especially at those favoured locations that had attracted high-quality FDI (Tang & Zhang, 2016; Teixeira & Fortuna, 2010). The growing complexity of technology dramatically increases the importance of transaction and coordination activities, which require even more physical INF. Governments often play an important role in improving INF on a time-bound basis. Therefore, one of the benefits from clustering is the improvement and establishment of specialized INF with the cooperation of the host government and industry associations (Porter, 1990). Chong and Khalifah (2019) examined the Malaysian manufacturing industries’ participation in international production and sharing. They found that outsourcing and processing trade activities and multinational enterprises’ intra-firm trade motivate bilateral production sharing exports.

Another challenge in establishing a high-technology firm is the accessibility to finance (Hoffman et al., 1998; Mason & Harrison, 2004). New high technology carries the perception of a high degree of risk to many (Deakins & Philpott, 1994). The intangible nature of R&D investment prevents firms from sourcing through conventional bank funding. Consequently, business angels, venture capital and equity markets become the main external funding sources for new technological ventures (Chorev & Anderson, 2006). A case study carried out by Carpenter and Petersen (2002) involved a group of newly established high-technology firms in the USA to prove that debt financing is generally insignificant, while equity financing significantly changes the size of the firm. It was further argued that the American high-technology firms expanded faster than the European high-technology firms, due to the fact that Americans had easily accessible equity markets. Hence, a powerful financial market would have deepened the effect in the development of high-technology industries.

The gains from international trade have always been understated. It is rational to say that trade opens new opportunities for growth, especially for developing countries. The idea is that imports often embody innovations that are unavailable to local economies (Schneider, 2005). An argument is made here to show evidence that the absorptive capability of the host country is the key explanatory variable for capturing opportunities presented by foreign affiliates. Absorptive capability is defined here as the ability to exploit internal, and especially external, knowledge that is embodied in intangible assets, which are recognized as the key drivers of enterprise expansion and performance (Cohen & Levinthal, 1990; Harris & Li, 2009). Following their arguments, it is rational to conclude that a firm should possess sufficient resources and capabilities in order to captivate new knowledge to overcome the initial (sunk) costs of competing in international markets to facilitate foreign competition and, hence, face a dual challenge of defeating rigidities and winning on novel knowledge (Eriksson & Johanson, 1997).

The central idea here is that a nation must invest in human resources to accumulate human capital (HC) in order to sustain economic growth. HC refers to the accumulated stock of skills and talents, acquired through formal and informal education, which should manifest in the educated and skilled workforce of the nation (Harbi et al., 2009). HC is a major absorptive capability in the production of knowledge or ideas—both are vital for the survival of high-technology enterprises. Furthermore, R&D investment only works with people capable of research. Without proper education and training on human resources, these processes of knowledge or technology transfer and creation, whether reactive or proactive, would not be possible (Haddad & Harrison, 1993). Xu (2000) reported the importance of adequate HC to manage technology and productivity spillovers, corresponding to the idea of the existence of minimum threshold in HC for growth by Borensztein et al. (1998). Furthermore, scientists and engineers should cooperate with universities and high-technology enterprises to construct a cycle of absorbing, modifying, creating and transferring new technologies (Romijn & Albaladejo, 2002). Shin et al. (2019) found that R&D activities of both small and medium enterprises (SMEs) and large-scale enterprises (LSEs) were active from 2004 to 2014, but the SMEs, in particular, took a great stride in their patenting and innovation activities. They also found that the more companies engaged in export activities, the more actively patents grew.

Apart from HC as AC, a new proposition on the index of economic freedom (EF) emerged. EF can be simply defined as ‘the absence of government coercion or constraint on the production, distribution, or consumption of goods and services beyond the extent necessary for citizens to protect and maintain liberty itself’ (Heritage Foundation, 2004). Economists agreed on the fact that freedom to select and supply resources, competition in business, free trade with the rest of the world, secure property rights, etc., are great additions to economic development (Azman-Saini et al., 2010). Interestingly, Islam (1996) suggested that EF influences per capita income positively for low-income countries, whereas it influences growth only for high-income countries. This study employs the same Fraser Institute’s EF index as AC to economic growth on an emerging economy, namely Malaysia. EF enhances a firm’s (or country) ability to absorb and internalize new technology from MNCs. Freedom of exchange across borders will help local firms to penetrate into the international arena for exporting purposes as well (Aitken et al., 1997), hence, promoting high technology trade.

This study aims to establish the role of trade openness (TO) as part of AC. Limited empirical studies have emerged from this point of view. Previous literature has supported the evidence of AC developing from transmission of knowledge and technology from superior countries (Martin & Bell, 2006; Schneider, 2005). Without TO, this knowledge transmission would have failed. In addition, absorptive capability is motivated when more spending on R&D successfully enhances the competitiveness of high-technology products. With profit-seeking firms (either local firms or foreign firms investing in the country) observing improvement in HTRD due to growth in competitiveness in product markets, firms are more interested to invest in R&D. Along with TO, HTRD will continue to surge (Aghion et al., 2005; Tingvall & Poldahl, 2006). This argument is supported by Brandt et al. (2017) as they discover Chinese industries that have lower tariffs tend to become more competitive. They also demonstrated the importance of a more liberal environment for inward FDI to further facilitate knowledge and technology transmission. The link of TO, knowledge transmission and enhancement in competitiveness of firms even preceded market reforms. It can be argued that a developing country, like Malaysia, will have to facilitate the flow of knowledge and technology transfer even before developing market reforms. These transfers occur among the R&D invested in the country and among FDI. As MNCs are in favour of doing business in the country, this study hypothesizes the role played by TO as a component of AC.

Building on the literature, this study suggests that the availability of factor conditions would more likely influence HTRD positively. Therefore:

H₁: The greater the level of R&D, the higher its HTRD.

H₂: The greater the level of FDI, the higher its HTRD.

H₃: The greater the availability of INF, the higher its HTRD.

H₄: The greater the financial development (FD), the higher its HTRD.

Establishing the supporting system that enhances the factor creation process is crucial. A relatively important absorptive capability that enhances factor creation is the upgrading of HC. High-technology firms require an adequate stock of technically qualified manpower to absorb, modify, create and transfer new technological information (Romijn & Albaladejo, 2002). Technological information comes from sources of FDI as they invest and manufacture in developing countries. Once again, technically qualified HC must be available to capture and absorb technological information (Ahmed, 2012; Root, 1992). Likewise, EF enhances a country’s ability to absorb and internalize new technology from MNCs (FDI) and contribute to the performance of high-technology production and trade (Aitken et al., 1997). Moreover, domestic innovation (R&D) activities can be improved as new technology information is allowed to flow freely across borders, thereby increasing the level of competitiveness of high-technology products (trade) (Brandt et al., 2017). As foreign trade serves as a carrier of knowledge, a country that is more open to technology imports derives greater benefits from foreign innovation (Coe et al., 1997). FDI from these firms is considered a potential channel, which provides access to advanced technologies available on the global marketplace (Zhu & Jeon, 2007).

Building on the literature, this study suggests that the presence of AC would more likely positively influence HTRD directly. Therefore:

H₅: The greater the level of AC, the higher its HTRD.

Empirical Model and Econometric Methodology

Based on the aforementioned literature discussion, the empirical model is as follows:

H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} F D I_{t} + β_{3} I N F_{t} + β_{4} F D_{t} + μ_{t}

(1)

The absorptive capacity is incorporated with Equation (1) as follows:

H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} A C_{t} + β_{3} I N F_{t} + β_{4} F D I_{t} + β_{5} F D_{t} + μ_{t}

(2)

As AC refers to HC, TO and EF, a modified version of Brandt et al. (2017) on interaction between (R&D) and FDI with AC, can be built.

Hence, the empirical models of interaction term between R&D and three AC variables are presented as:

\begin{matrix} H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} H C_{t} + β_{3} {(R D \times H C)}_{t} \\ + β_{4} I N F_{t} + β_{5} F D I_{t} + β_{6} F D_{t} + μ_{t} \end{matrix}

(3)

\begin{matrix} H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} T O_{t} + β_{3} {(R D \times T O)}_{t} \\ + β_{4} I N F_{t} + β_{5} F D I_{t} + β_{6} F D_{t} + μ_{t} \end{matrix}

(4)

\begin{matrix} H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} E F_{t} + β_{3} {(R D \times E F)}_{t} \\ + β_{4} I N F_{t} + β_{5} F D I_{t} + β_{6} F D_{t} + μ_{t} \end{matrix}

(5)

The empirical models of interaction term between FDI and three AC variables are as follows:

\begin{matrix} H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} I N F_{t} + β_{3} F D I_{t} + β_{4} H C_{t} \\ + β_{5} {(F D I \times H C)}_{t} + β_{6} F D_{t} + μ_{t} \end{matrix}

(6)

\begin{matrix} H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} I N F_{t} + β_{3} F D I_{t} + β_{4} T O_{t} \\ + β_{5} {(F D I \times T O)}_{t} + β_{6} F D_{t} + μ_{t} \end{matrix}

(7)

\begin{matrix} H T R D_{t} = β_{0} + β_{1} R D_{t} + β_{2} I N F_{t} + β_{3} F D I_{t} + β_{4} E F_{t} \\ + β_{5} {(F D I \times E F)}_{t} + β_{6} F D_{t} + μ_{t} \end{matrix}

(8)

If the interaction between R&D and AC (i.e., RD × TO) is statistically significant, it signifies the role of AC in boosting the effect of R&D. Likewise, if the interaction between FDI and AC (i.e., FDI × TO) is statistically significant, it signifies the role of AC in boosting the effect of FDI (Teixeira & Fortuna, 2010). Thus, the importance of having AC will not be neglected.

Data Description

All data for R&D, INF, FD, HC and TO are obtained from the WDI and the World Bank. R&D measures the expenditures for R&D based on current and capital expenditures (both public and private) on creative work undertaken systematically to increase knowledge and the use of knowledge for new applicants. It is taken as a sum of expenditure on R&D. FD or domestic credit to the private sector refers to financial resources provided to the private sector by financial corporations, such as through loans, purchases of non-equity securities and trade credits, and other accounts receivable that establish a claim for repayment. It is expressed in terms of percentage of GDP. INF shows the investment in transport and telecom projects with private participation and covers INF projects in transport and telecommunications that have reached financial closure. HC is measured using total enrolment in secondary education and expressed as a percentage of population of official secondary education rate. TO is compiled using total trade as a percentage of GDP (Squalli & Wilson, 2011).

HTRD shows the sum of exports and imports of high-technology products (Squalli & Wilson, 2011). FDI refers to the net inward flows, which are expressed in USD million. HTRD data are obtained from the Malaysian Department of Statistics, while data on FDI are obtained from the United Nations Conference on Trade and Development (UNCTAD), and EF data are obtained from Fraser Institute. EF measures the index of EF in five broad areas, namely (a) size of government: expenditures, taxes and enterprises; (b) legal structure and security of property rights; (c) access to sound money; (d) freedom to trade internationally; and (e) regulation of credit, labour and business. The quarterly data cover the sample period between 1990 and 2015. All data have been transformed into logarithm form before analysis.

Autoregressive Distributed Lag Approach

Pesaran et al. (2001) advocated for the use of the autoregressive distributed lag (ARDL) model by suggesting that once the order of ARDL has been identified, estimation and identification can be estimated by using ordinary least square (OLS). The ARDL model also captures the effects from lagged independent and dependent variables, and by including sufficient number of lags, to eliminate serial correlation in the errors (Hill et al., 2012). Furthermore, the asymptotic distribution of the F-statistics is non-standard under the null hypothesis of no cointegration between the examined variables, irrespective of whether the explanatory variables are purely I(0) or I(1), or mutually cointegrated. In other words, bounds test allows for the unsynchronized order of integration between interested variables. Hence, it has an edge of not requiring a precise identification of order of the underlying data. This multivariate cointegration procedure also deems appropriate for the model as it caters better smaller sample size properties. Therefore, an ARDL testing approach is selected to estimate the model specifications.³

Assuming that the linear bounds test leads to the conclusion of cointegration and there is no non-linear relationship, we can estimate the long-run cointegration relationship among the variables based on the uniform lag length proposed by Pesaran et al. (2001) (p, p, p, p, p, p) as follows:

\begin{matrix} H T R D_{t} = β_{0} + \sum_{i = 1}^{p} β_{1} H T R D_{t - i} + \sum_{j = 0}^{p} γ_{j} Δ R D_{t - j} + \sum_{k = 0}^{p} δ_{k} Δ A C_{t - k} \\ + \sum_{1 = 0}^{p} λ_{1} Δ I N F_{t - 1} + \sum_{m = 0}^{p} ϕ_{m} Δ F D I_{t - m} + \sum_{n = 0}^{p} η_{n} Δ F D_{t - n} + θ_{0} H T R D_{t - 1} \\ + θ_{1} R D_{t - 1} + θ_{2} A C_{t - 1} + θ_{3} I N F_{t - 1} + θ_{4} F D I_{t - 1} + θ_{5} F D_{t - 1} + e_{t} \end{matrix}

(9)

The uniform lag length is subject to serial correlation test (if there is a serial correlation problem, then the general-to-specific approach is used to select different lag lengths (p, q, r, s, t, u). For example, let us say ARDL (1, 1, 1, 1, 1, 1) is the optimal lagged model, then the equation is as follows:

\begin{matrix} H T R D_{t} = μ_{i} + β H T R D_{t - 1} + δ_{1} R D_{t} + δ_{2} R D_{t - 1} + δ_{3} A C_{t} + δ_{4} A C_{t - 1} + \\ δ_{5} I N F_{t} + δ_{5} I N F_{t - 1} + δ_{6} F D I_{t} + δ_{7} F D I_{t - 1} + δ_{8} F D_{t} + δ_{9} F D_{t - 1} + ε_{t} \\ δ_{5} I N F_{t - 1} + δ_{6} F D I_{t} + δ_{7} F D I_{t - 1} + δ_{8} F D_{t} + δ_{9} F D_{t - 1} + ε_{t} \end{matrix}

(10)

and the short-run error correlation model (ECM) equation is represented as follows:

\begin{matrix} Δ H T R D_{t} = β_{0} + \sum_{i = 1}^{p} β_{i} Δ H T R D_{t - i} + \sum_{j = 0}^{p} γ_{j} Δ R D_{t - j} + \sum_{k = 0}^{r} δ_{k} Δ A C_{t - k} \\ + \sum_{m = 0}^{s} η_{m} Δ I N F_{- m} + \sum_{n = 0}^{v} θ_{n} Δ F D I_{t - n} \sum_{i = 1}^{w} τ_{o} Δ F D_{t - o} + φ z_{t - 1} + e_{t} \end{matrix}

(11)

where z_t_–1 = (HTRD_t–1 – α₀ – α₁RD_t–1 – α₂AC_t–1 – α₃INF_t–1 – α₄FDI_t–1 – α₃FD_t–1) or the error-correction term (ECT) and the αs are the OLS estimates of the αs in Equation (11). The f in Equation (11) is the short-run equation that contains the ECT that measures the speed of adjustment of the short-run to long-run equilibrium. p, q, r, s, v and w are the optimal lagged lengths, selected using the Schwarz Bayesian criterion (SBC). Based on this ARDL (1, 1, 1, 1, 1, 1) model, we can compute the long-run coefficients of the determinants⁴:

\begin{matrix} ϕ_{R D} = \frac{γ_{1} + γ_{2}}{1 - β_{i}}; ϕ_{A C} = \frac{δ_{1} + δ_{2}}{1 - β_{i}}; ϕ_{I N F} = \frac{η_{1} + η_{2}}{1 - β_{i}}; \\ ϕ_{F D I} = \frac{θ_{1} + θ_{2}}{1 - β_{i}}; ϕ_{F D} = \frac{τ_{1} + τ_{2}}{1 - β_{i}} \end{matrix}

Besides the ARDL model, various statistical methods have been used in testing cointegration, especially under the condition of non-stationary phenomenon to avoid spurious correlations in recent decades. Another method used among time-series cointegration analysis is the fully modified ordinary least squares (FMOLS). Philips and Hansen (1990) applied a two-part transformation procedure to remove the asymptotic bias terms, which require an estimation of long-run variance matrices. FMOLS is designed to account for serial correlation effects and endogeneity among regressors that result from the existence of a cointegrating relationship. The difference between FMOLS and ARDL is that it does not involve any stationary or cointegration hypothesis testing. As an alternative, Philips and Hansen (1990) focused more on the estimated coefficient bias rather than the existence of stationary properties in the error term. With different focuses on the band, this method would be suitable as an option for robustness inspection in the study.

III. Empirical Results

Table 1 demonstrates the results of the unit root test for all the variables used to answer the first objective of the study. Augmented Dickey–Fuller test is used to recognize the stationarity of the data. Following Pesaran and Pesaran (1997), HTRD, being the endogenous variable, has to be stationary at the first difference level. The result from Table 1 suggests that HTRD is stationary at first difference I(1) level at the 1% and 5% significance level. All other variables, namely RD, HC, RDHC, FDIHC, TO, RDTO, EF, RDEF, INF, FDI and FD, have to be ascertained at level or the first difference altitude. Conclusively, all variables fulfil the stationary condition at first difference I(1) at the 1% significance level. Therefore, the empirical models support the presence of a unit root in the level of all variables and the absence of any unit root after the first differencing. The process continues with bounds test.

Table 1.
Results of the Unit Root Test

Augmented Dickey–Fuller (ADF) Test

Variables Level
First Difference

Intercept Trend and Intercept Intercept Trend and Intercept

HTRD –2.605808*
(8) –1.786170
(8) –4.311698*
(3) –3.518148
(7)

RD –0.470150
(0) –2.670484
(0) –10.49257*
(0) –10.44232*
(0)

FDI –2.146831
(8) –2.485684
(8) –6.048761*
(7) –6.042710*
(7)

HC –1.374551
(0) –1.829747
(0) –10.17736*
(0) –10.15213*
(0)

RDHC –1.343567
(0) –1.406002
(0) –10.01292*
(0) –10.09381*
(0)

FDIHC –3.499769*
(0) –4.055738*
(0) –10.02424*
(0) –10.39800*
(0)

TO –1.648714
(4) –1.569296
(4) –3.909650*
(3) –3.927328
(3)

RDTO –1.522028
(0) –1.400890
(0) –10.01116*
(0) –10.18895*
(0)

FDITO –3.931173*
(0) –3.865524
(0) –10.02007*
(0) –10.17900*
(0)

EF –1.524225
(0) –1.273927
(0) –10.00087*
(0) –10.02648*
(0)

RDEF –2.325117
(0) –2.305724
(0) –10.00642*
(0) –9.978761*
(0)

FDIEF –2.189357
(4) –2.365999
(4) –8.650271*
(3) –8.615519*
(3)

INF –2.449439
(0) –2.405647
(0) –10.00348*
(0) –9.968524*
(0)

FD –2.942100
(4) –3.030531
(4) –3.905307*
(3) –4.061574***
(3)

Augmented Dickey–Fuller (ADF) Test
HTRD	–2.605808* (8)	–1.786170 (8)	–4.311698*** (3)	–3.518148** (7)
RD	–0.470150 (0)	–2.670484 (0)	–10.49257*** (0)	–10.44232*** (0)
FDI	–2.146831 (8)	–2.485684 (8)	–6.048761*** (7)	–6.042710*** (7)
HC	–1.374551 (0)	–1.829747 (0)	–10.17736*** (0)	–10.15213*** (0)
RDHC	–1.343567 (0)	–1.406002 (0)	–10.01292*** (0)	–10.09381*** (0)
FDIHC	–3.499769*** (0)	–4.055738*** (0)	–10.02424*** (0)	–10.39800*** (0)
TO	–1.648714 (4)	–1.569296 (4)	–3.909650*** (3)	–3.927328** (3)
RDTO	–1.522028 (0)	–1.400890 (0)	–10.01116*** (0)	–10.18895*** (0)
FDITO	–3.931173*** (0)	–3.865524** (0)	–10.02007*** (0)	–10.17900*** (0)
EF	–1.524225 (0)	–1.273927 (0)	–10.00087*** (0)	–10.02648*** (0)
RDEF	–2.325117 (0)	–2.305724 (0)	–10.00642*** (0)	–9.978761*** (0)
FDIEF	–2.189357 (4)	–2.365999 (4)	–8.650271*** (3)	–8.615519*** (3)
INF	–2.449439 (0)	–2.405647 (0)	–10.00348*** (0)	–9.968524*** (0)
FD	–2.942100** (4)	–3.030531 (4)	–3.905307*** (3)	–4.061574*** (3)

Source: The authors.

Note: (1) *** Indicates the rejection of null hypothesis of non-stationarity at the 1% level of significance.

(2) ** Indicates the rejection of null hypothesis of non-stationarity at the 5% level of significance.

(3) * Indicates the rejection of null hypothesis of non-stationarity at the 10% level of significance.

(4) The figure in parentheses () refers to the optimal selected lag length.

Table 2 reports the bounds cointegration of three models where the critical values in Pesaran et al. (2001) are used to compare with the computed F-statistics from the test. As F-statistics of all models are larger than the critical values of the upper bound, null hypothesis—non-existence of long-run relationship—is rejected at the 1% level of significance. Thus, the result suggests the existence of a steady-state long-run relationship among HTRD, RD, INF, FDI and FD. After examining for bounds cointegration test, the process continues into the estimation of ARDL level relation to find out the long-run and short-run dynamic relationships.

Table 2.

Result for Bounds Test for Cointegration Analysis

Models		F-statistic
Model 1: HTRD = f (RD, HC, INF, FDI, FD)		8.5811***
Model 2: HTRD = f (RD, TO, INF, FDI, FD)		8.0345***
Model 3: HTRD = f (RD, EF, INF, FDI, FD)		10.0268***
	k = 6, n = 100
Pesaran et al. (2001) Critical Value	Lower Bound	Upper Bound
10%	2.12	3.23
5%	2.45	3.61
1%	3.15	4.43

Source: The authors.

Notes: *** Denote significance at the 1% level.

Table 3 presents the results of long-run coefficients using ARDL and FMOLS estimations and short-run ARDL dynamic coefficients for three models. These three models include the interaction terms: (RD × HC), (RD × TO) and (RD × EF), respectively. Under Model 1, that is (RD × HC), only FDI serves a significant variable determinant of HTRD in the long run. The error correction coefficient is negative and statistically significant, which is in line with the bounds long-run cointegration result. The ERT is 6.4%, which implies that any short-run deviation will take about 15.62 quarters or 3.9 years to adjust back to a long-run equilibrium path or speed of convergence.⁵ Meanwhile, for Model 2, that is (RD × TO), the long-run coefficient results using ARDL reveal that TO and FDI have a significant impact on HTRD at the 1% level of significance, whereas RDTO, FD and INF have a significant impact at the 10% level of significance. The result of FMOLS estimation is also consistent with the major findings of ARDL model, suggesting robustness in the findings. The error correction coefficient is at 6.7%, which is considered as a low speed of convergence or 15 quarters or 3.73 years to move back to the long-run equilibrium if they have any short-run deviation. The analysis of the model with interaction between R&D and TO (RD × TO) is more likely to be selected under all circumstances. This model demonstrates promising results as interaction with AC is taken into account (RDTO). Additionally, the long-run coefficient results for Model 3 (RD × EF), using ARDL approach, indicates that only FDI shows significant influence on HTRD. The FMOLS long-run estimation result is not the same with the ARDL findings; this is not surprising since both estimators are different. The error correction coefficient is at 19% but significant at weak level (10%), where any short-run deviation will take about 5.26 quarters or 1.32 years to adjust back to the long-run equilibrium.⁶ In short, the long-run estimation results imply that FDI and TO are two important variables that determine HTRD in the long run.

Table 3.

Estimated Long-run Coefficient and Short-run Dynamic ECT Model

Dependent Variable: HTRD
Variable	Model 1: (RD × HC)		Model 2: (RD × TO)		Model 3: (RD × EF)
	Long-run Coefficient
	ARDL (4, 1, 0, 0, 1, 4, 0)	FMOLS	ARDL (4, 0, 1, 1, 1, 1, 1)	FMOLS	ARDL (4, 1, 0, 1, 1, 3, 0)	FMOLS
RD	0.1768	0.8155***	0.0253	0.7787	0.0102	0.8687***
HC	1.4071	1.2024***	–	–	–	–
TO	–	–	1.6412***	0.5705***	–	–
EF	–	–	–	–	0.8976	–1.4442***
RDHC	2.1108	–4.7374***	–	–	–	–
RDTO	–	–	1.8023*	0.7082***	–	–
RDEF	–	–	–	–	–7.0716	–1.3431
FDI	0.3920***	0.1246***	0.3922***	0.0900***	0.4417**	0.1024***
FD	–0.3001	0.5867***	1.2028*	0.2949***	–0.2758	0.7936***
INF	0.0669	0.2308***	0.2543*	0.0823**	0.1667	–0.0078
Constant	–0.4375	–1.5772	–2.3941	6.0017	0.4651	10.9375
Short-run Dynamic ECT Model
Variable	Coefficient		Coefficient		Coefficient
D(RD)	0.1304**		–0.0017		0.1677***
D(HC)	0.0897		–		–
D(TO)	–		0.5216***		–
D(EF)	–		–		0.0489
D(RDHC)	0.1347		–		–
D(RDTO)	–		0.3808***		–
D(RDEF)	–		–		1.1900
D(FDI)	0.0506***		0.0576***		0.0502***
D(FD)	–0.1953		0.0215		–0.1792
D(FD(-1))	–0.0065		–		–0.2535
D(FD(-2))	–0.3085		–		–0.2585
D(FD(-3))	–0.2546		–		–
D(INF)	0.0043		0.0532***		0.0545**
ECT(-1)	–0.0638**		–0.0677***		–0.1901*

Source: The authors.

Notes: *, ** and *** Denote significance at the 10%, 5% and 1% levels, respectively. The HTRD short-run coefficients are not reported to save space.

To verify the models’ diagnostic checks, this study presents the autocorrelation tests, normality test and Ramsey test in Table 4. As the level of significance (p-value) for all statistics is higher than 0.05, at 95% confidence level, the three models fulfil the econometric properties. Figure 1 presents the results of cumulative sum (CUSUM) stability test, and all three models are stable.

Table 4.

Results of the ARDL Diagnostic Checks of Models 1–3

Model	Test Type and p-Values
Model	Serial Correlation LM test	Ramsey RESET Test	Residual Normality Test
Model 1: RD × HC	0.1202	0.1074	0.6128
Model 2: RD × TO	0.2003	0.3002	0.4178
Model 3: RD × EF	0.1102	0.1001	0.0946

Source: The authors.

Figure 1.

Source: The authors.

In comparison to all three models, only when R&D interacts with TO, (RD × TO) serves the best purpose as an absorptive capability that could fire up the catching up process—TO is significantly influencing in both the short run and long run. When TO is present, knowledge and technology transfer occurs through the process of import. Imports of high-technology product made new technologies available in the local market, thereby facilitating the process of improvement in R&D of the particular technology (Schneider, 2005). Once an addition to the existing technology is available through local R&D, the new product can create a high-impact growth to the export of HTRD. Note that under Model 2 (RD × TO), RD is not enough to show significant results, but the interaction of R&D and TO (RDTO) shows significant results. Despite heavy expenditure spending on R&D in the country, it is not enough to promote growth unless it includes AC like TO. Without a secured level of AC, the projection of local R&D could not be magnified (Cohen & Levinthal, 1990). HTRD cannot be successful with only heavy expenditure on R&D in the case of developing countries like Malaysia. As suggested by Kim (1980), developing countries must possess the ability to facilitate and develop technological capabilities. Without these capabilities, expenditures on R&D would not be able to project high-value growth to the economy.

A point to note is that FDI also has a significant impact in influencing HTRD in both the short run and long run in all models. This validates the findings from the literature that FDI aids technology transfer and trade to developing countries (Root, 1992; Seyoum, 2004). With necessary technical support (i.e., AC), firms have integrated with foreign affiliates to boost technological transfer and HTRD. Therefore, the next subsection will extend the interest into interaction of FDI with different AC.

Under model (RD × TO), INF also has a positive and significant influence on HTRD. This proves that the development in transaction and coordination activities positively correlated with the development of HTRD. The absence of modern INF can create problems in the introduction and adoption of modern technologies and competition in international markets (Porter, 1990; Seyoum, 2004). Hence, this study confirms INF as one important determinant of HTRD. Another determinant that is positively significant to influence HTRD in the long run is FD. However, this determinant is not significant in the short run. As pointed out by Deakins and Philpott (1994), there is a degree of risk on new high technology and asymmetric information between firms and the financial supplier. Therefore, limited financial possibilities would occur in the short run (Harbi et al., 2009). In the long run, FD maintains as a determinant of HTRD as accessibility to finance is much more available after long evaluations and understanding of new high technology.

The analyses continue with demonstrating the effect of FDI and interaction with three similar ACs, namely HC, TO and EF. Table 5 presents the results of bounds cointegration test and the findings suggest the existence of a steady-state long-run relationship among independent and dependent variables in Models 4–6. After examining the bounds cointegration test, the process continues to estimate the ARDL level relation of log-run and short-run relationships.

Table 5.

Result for Bounds Test for Cointegration Analysis

Models		F-statistic
Model 4: HTRD = f (RD, INF, FDI, HC, FD)		10.3146***
Model 5: HTRD = f (RD, INF, FDI, TO, FD)		6.7989***
Model 6: HTRD = f (RD, INF, FDI, EF, FD)		9.5523***
	k = 6, n = 100
Pesaran et al. (2001) Critical Value	Lower Bound	Upper Bound
10%	2.12	3.23
5%	2.45	3.61
1%	3.15	4.43

Source: The authors.

Notes: *, ** and *** Denote significance at the 10%, 5% and 1% levels, respectively.

Table 6 summarizes the empirical results of long-run coefficients using ARDL and FMOLS and the short-run coefficients using ARDL for all three models. As for Model 4 (FDI × HC), the findings from ARDL demonstrate that only FDI is the statistically significant determinant of HTRD in the long run. As the findings from the long-run coefficients of both ARDL and FMOLS are different, this study needs to interpret the results with caution, since different estimators yield different findings. The error-correction coefficient is recorded at 9.4% in Model 4, which shows an average speed of convergence or 10.64 quarters or 2.66 years. Based on Model 5 (FDI × TO), all three variables of interest, namely FDI, TO and FDITO, reveal a significant impact on HTRD in the long run. Comparing with the results from FMOLS, results from the ARDL model are to be robust in this case. The speed of convergence is considerably slower among the three models and is estimated at 6.4% or about 15.63 quarters or 3.9 years to adjust back to the long-run equilibrium. For Model 6 (FDI × EF), the findings from ARDL proved that FDI, FDIEF and INF have a significant impact on HTRD in the long run. However, the expected sign of FDIEF is considered as the reverse of what is expected by previous literature. Previous studies by Doucouliagos and Ulubasoglu (2006), Islam (1996) and Scully and Slottje (1991) proved there is a positive relationship between EF and economic growth. In this case, EF is negatively associated with economic growth. Thus, this study goes along with the empirical evidences from de Hann and Siermann (1998) that different measurements of EF will alter the estimated effect. However, results from FMOLS are dissimilar with results of ARDL, indicating that different estimators have dissimilar findings in this case. The error-correction coefficient is at 6.9%, at an average speed of convergence or takes about 14.5 quarters or 3.6 years to adjust towards the long-run equilibrium path. The empirical findings of Table 6 are in line with Table 3, where FDI and TO are also significant determinants of HTRD in the long run.

Table 6.

Estimated Long-run Coefficient and Short-run Dynamic ECT Model

Dependent Variable: HTRD
Variable	Model 4: (FDI × HC)			Model 5: (FDI × TO)		Model 6: (FDI × EF)
	Long-run Coefficient
	ARDL (4, 1, 1, 0, 4, 4, 0)		FMOLS	ARDL (4, 0, 1, 1, 0, 0, 1)	FMOLS	ARDL (4, 0, 1, 0, 0, 3, 0)	FMOLS
RD	0.1422		0.8888***	0.0363	0.7775	0.3295	0.8821***
FDI	0.3399***		0.0657***	0.4016***	0.0913***	0.3676***	0.0972***
HC	2.2502		0.5163***	–	–	–	–
TO	–		–	1.1251**	0.4879***	–	–
EF	–		–	–	–	2.9396	–1.2077***
FDIHC	1.8331		–1.9237***	–	–	–	–
FDITO	–		–	1.3272*	0.0104***	–	–
FDIEF	–		–	–	–	–12.6824***	–0.5456
FD	–0.5745		0.4896***	–0.5728	0.4331***	0.0847	0.8175***
INF	0.0595		0.0113	0.1321	0.0555***	0.4048**	–0.0002
Constant	–1.5570		7.8207	–0.4001	6.2924	–8.4715	10.3207
Short-run Dynamic ECT Model
Variable		Coefficient		Coefficient		Coefficient
D(RD)		0.1283**		–0.0023		0.0226
D(FDI)		0.05933***		0.0548***		0.0443***
D(HC)		0.2112*		–		–
D(TO)		–		0.4634***		–
D(EF)		–		–		0.2017*
D(FDIHC)		0.8175		–		–
D(FDIHC(-1))		–0.2535		–		–
D(FDIHC(-2))		–0.2585		–		–
D(FDIHC(-3))		0.5998		–		–
D(FDITO)		–		0.0843*		–
D(FDIEF)		–		–		–0.8704***
D(FD)		–0.1906		–0.0364		–0.1937
D(FD(-1))		–0.0065		–		0.0091
D(FD(-2))		–0.3085		–		–0.2562
D(FD(-3))		–0.2233		–		–
D(INF)		0.0056		0.0463***		0.0278***
ECT(-1)		–0.0938***		–0.0635***		–0.0686***

Source: The authors.

Notes: *, ** and *** Denote significance at the 10%, 5% and 1% levels, respectively.

To ensure the ARDL models were well specified, this study performed autocorrelation test, normality test and Ramsey test, and results are presented in Table 7. As the levels of significance for all statistics were higher than 0.05, in 95% confidence level, the three models were well defined and passed the diagnostic checks. Figure 2 presents the results of CUSUM stability test, and all three models are stable.

Table 7.

Results of the Diagnostic Checks of Models 4–6

Model	Test Type and p-Values
Model	Serial Correlation LM Test	Ramsey RESET Test	Residual Normality Test
Model 4: FDI × HC	0.1002	0.1262	0.1128
Model 5: FDI × TO	0.2031	0.2013	0.2191
Model 6: FDI × EF	0.1201	0.2082	0.0895

Source: The authors.

Figure 2.

Source: The authors.

Nonetheless, still TO serves the best purpose of AC in the expansion of HTRD. It significantly influences during both the short run and long run with FDI. FDI with the interaction of HC (FDIHC) does not show significant influence on HTRD, and interaction with EF (FDIEF) shows an adverse impact on HTRD. Therefore, Model 5 (FDI × TO) is still a favourable model as three main variables of interest, which are FDI, TO and interaction between FDI and TO (FDITO), show significant influence during both the short run and long run. It is interesting to note that all three models do not support the notion that R&D will help the high-technology industry under the presence of FDI. This is due to the levels of technology and knowledge that had been transferred to Malaysia, through FDI, which had a greater impact than the local R&D effort (Root, 1992; Seyoum, 2004)—local R&D has an impact in the long run only, and the expenditure of R&D needs to increase. As the TO widens the channel of knowledge and technology diffusion, the benefits of FDI is magnified, promoting the performance in HTRD eventually. Developing economies like Malaysia still has to put greater effort into attracting quality FDI to promote the ladder of convergence. Another determinant that significantly influences HTRD in the short run under Model 5 (FDI × TO) is INF (INF). Once again, the importance of existing INF has been stressed to improve the transaction and coordination activities in the execution of HTRD. The presence of modern physical INF, such as transportation, computers and telecommunications, eases the process of introduction and adoption of new technology (Seyoum, 2004).

IV. Conclusions

This study examined the determinants of HTRD and the appropriate absorptive capability, in enhancing economic growth in Malaysia, using quarterly data from 1990 to 2015. The empirical findings revealed that FDI, FD and INF were vital to develop a successful HTRD. TO was an important absorptive capability to boost the benefits of R&D. In addition, there was long-run convergence among the variables used in the analysis. Based on the empirical results, it was undeniably important to practice TO in the country. Malaysia should promote more trade induction facilities like tariff exemption and trade agreements that benefit the trade environment in the country and others. General trade policies should focus on increasing market access for export and greater global integration as it was statistically significant in magnifying the benefits from R&D and, eventually, HTRD. Malaysia, an emerging market, should possess the ability to facilitate and develop technological capabilities—not just increasing the spending on R&D. Furthermore, policies to enhance comparative advantage in HTRD will increase the level of competitiveness in international market, thereby leading to economic growth.

Apart from TO, FDI also shows significant empirical results in both the short run and long run. Policies that attract quality FDI should not be ignored. The establishment of the Iskandar Development Region in the southern part of Malaysia is an example of creating an investor-friendly environment. Specifically in the case of developing countries, most firms are basically government-owned or state-owned enterprises where efficiency remains doubtful. The reduction of entry barriers and allowing multinational enterprises to enter domestic markets via FDI would promote knowledge and technology spillover. These interactions with local researchers will enhance the creation of new technology, which is substantial for the creation of high-impact economic growth for the country.

Besides FDI, the presence of INF and FD is also substantial in the performance of HTRD. The use of clustering (clustering is indeed the word often used) in industry associations and to include local governments to cooperate and establish their own specialized INF. With the effects from clustering such as science parks or entrepreneurial centres, regional development benefits from cost-saving set-ups and positive externalities shared among community members. As financial access is limited to high-technology industry, the government has to establish financial mechanisms that are specific to high-technology ventures and are independent of banking power. With the existence of matured financial systems, it is more likely to attract new high-technology firms and to expand existing firms’ production to increase the market share and competitiveness levels in international trade.

Footnotes

Acknowledgements

We would like to thank the editor for the useful suggestions on an earlier version of this article. Helpful comments and suggestions from the anonymous reviewers are greatly appreciated.

Declaration of Conflicting Interests

The authors declare no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We are grateful for the financial support from the Graduate Research Fellowship (GRF) research grant, Universiti Putra Malaysia for funding this work.

Notes

References

Abramovitz

(1986). Catching up, forging ahead, and falling behind. Journal of Economic History, 46(2), 385–406.

Aghion

, Bloom

, Blundell

, Griffith

& Howitt

(2005). Competition and innovation: An inverted U-relationship. The Quarterly Journal of Economics, 120(2), 701–728.

Ahmed

E. M.

(2012). Are the FDI inflow spillover effect on Malaysia’s economic growth input driven? Economic Modelling, 29, 1498–1504.

Aitken

, Hanson

, & Harrison

(1997). Spillovers, foreign investment, and export behavior. Journal of International Economics, 43, 103–132.

Alfaro

, Chanda

, Kalemli-Ozcan

, & Sayek

(2004). FDI and economic growth: The role of local financial markets. Journal of International Economics, 64, 89–112.

Archibugi

, & Pietrobelli

(2003). The globalisation of technology and its implications for developing countries: Windows of opportunity or further burden? Technological Forecasting and Social Change, 70, 861–883. 10.1016/S0040-1625(02)00409-2

Athreye

, & Cantwell

(2007). Creating competition? Globalisation and the emergence of new technology producers. Research Policy, 36, 209–226.

Azman-Saini

W. N. W.

, Baharumshah

A. Z.

, & Law

S. H.

(2010). Foreign direct investment, economic freedom and economic growth: International evidence. Economic Modeling, 27, 1079–1089.

Azman-Saini

W. N. W.

, Farhan

M. Z. M.

, Tee

C. -L.

, & Tun

Y. -L.

(2018). FDI inflows and R&D activity in developing countries. International Journal of Economics and Management, 12(S2), 509–521.

10.

Batten

J. A.

, & Vo

X. V.

(2009). An analysis of the relationship between foreign direct investment and economic growth. Applied Economics, 41, 1621–1641.

11.

Birkinshaw

, Hood

, & Jonsson

(1998). Building firm-specific advantages in multinational corporations: The role of subsidiary initiative. Strategic Management Journal, 19(3), 221–241.

12.

Brandt

, Van Biesebroeck

, Wang

& Zhang

(2017). WTO accession and performance of Chinese manufacturing firms. American Economic Review, 107(9), 2784–2820.

13.

Borensztein

, De Gregorio

, & Lee

J. W.

(1998). How does foreign direct investment affect economic growth. Journal of International Economics, 45, 115–135.

14.

Carolan

, Mora

, & Singh

(2013). Trade dynamics in the East Asian miracle: A time series analysis of US-East Asia commodity trade, 1962–1992. Millennial Asia, 4(1), 87–108.

15.

Cantwell

(1999). The globalization of technology: What remains of the product life cycle model. In Archibugi

, Howells

, & Michie

(Eds.), Innovative policy in a global economy (pp. 225–242). Cambridge University Press.

16.

Carpenter

R. E.

, & Petersen

B. C.

(2002). Capital market imperfections, high-tech investment, and new equity financing. Economic Journal, 112(477), 54–72.

17.

Cherif

, & Hasanov

(2015). The leap of the tiger: How Malaysia can escape the middle-income trap (IMF Working Paper, WP/15/131). International Monetary Fund.

18.

Chong

P. W.

, & Khalifah

N. A.

(2019). Export expansion and production sharing in Malaysian manufacturing. International Journal of Economics and Management, 13(1), 37–49.

19.

Chorev

, & Anderson

A. R.

(2006). An aesthetic rural environment: The Tefen model of rural development. The Journal of Rural Enterprise and Management, 3(1), 56–68.

20.

Chuang

W. B.

, & Lin

H. L.

(2010). Interdependence between overseas and domestic R&D activities: Evidence from Taiwanese multinationals. Asian Economic Journal, 24, 305–332.

21.

Coe

D. T.

, Helpman

, & Hoffmaister

A. W.

(1997). North–South R&D spillovers. Economic Journal, 107(440), 134–149.

22.

Cohen

W. M.

, & Levinthal

D. A.

(1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35(1), 128–152.

23.

Deakins

, & Philpott

(1994). Comparative European practices in the finance of new technology entrepreneurs: United Kingdom, Germany and Holland. In Oakey

R. P.

(Ed.), New technology-based firms in the 1990s. Paul Chapman Publishing.

24.

Doucouliagos

, & Ulubasoglu

M. A.

(2006). Economic freedom and economic growth: Does specification make a difference? European Journal of Political Economy, 22(1), 68–81.

25.

Eriksson

, & Johanson

(1997). Experiential knowledge and cost in the internalisation process. Journal of International Business Studies, 28, 337–360.

26.

Fagerberg

, & Godinho

M. M.

(2005). Innovation and catching-up. In Fagerberg

, Mowery

, & Nelson

(Eds.), The Oxford handbook of innovation. Oxford University Press.

27.

Falk

(2007). R&D spending in the high-tech sector and economic growth. Research in Economics, 61(3), 140–147.

28.

Falkowski

(2018). Competitiveness of the Baltic states in international high-technology goods trade. Comparative Economic Research, 21(1), 25–43.

29.

Fotros

M. H.

, & Ahmad

(2017). Impact of import and export of high technology industries on economic growth of Iran. Quarterly Journal of Quantitative Economics, 14(2), 1–24.

30.

Gamal

A. A. M.

, Rambeli

, Jalil

N. A.

, & Viswanathan

K. K.

(2019). A modified currency demand function and the Malaysian shadow economy: Evidence from ARDL bounds testing approach. Economic Analysis and Policy, 64, 266–281.

31.

Grossman

, & Helpman

(1991). Innovation and growth in the global economy. MIT Press.

32.

Guimon

, Chaminade

, Maggi

, & Salazar-Elena

J. C.

(2018). Policies to attract R&D-related FDI in small emerging countries: Aligning incentives with local linkages and absorptive capacities in Chile. Journal of International Management, 24, 165–178.

33.

Guo

, Bo

, Tong

, & Zhang

(2020). A paradoxical view of speed and quality on operational outcome: An empirical investigation of innovation in high-tech small and medium-sized enterprises. International Journal of Production Economics, 229, 107780.

34.

Haddad

, & Harrison

(1993). Are there positive spillovers from direct foreign investment? Evidence from panel data for Morocco. Journal of Development Economics, 42, 51–74.

35.

de Hann

, & Siermann

C. L. J.

(1998). Further evidence on the relationship between economic freedom and economic growth. Public Choice, 95, 363–380.

36.

Harbi

, Amamou

, & Anderson

A. R.

(2009). Establishing high-tech industry: The Tunisian ICT experience. Technovation, 29, 465–480.

37.

Harris

, & Li

Q. C.

(2009). Exporting, R&D, and absorptive capacity in UK establishments. Oxford Economic Papers, 61, 74–103.

38.

Heritage Foundation. (2004). Index of economic freedom. The Heritage Foundation and Dow Jones and Company.

39.

Hermes

, & Lensink

(2003). Foreign direct investment, financial development and economic growth. Journal of Development Studies, 40, 142–163.

40.

Hill

R. C.

, Griffiths

W. E.

, & Lim

G. C.

(2012). Principles of econometrics. John Wiley.

41.

Hoffman

, Parejo

, Bessant

, & Perren

(1998). Small firms, R&D, technology and innovation in the UK: A literature review. Technovation, 18(1), 39–55.

42.

Islam

(1996). Economic freedom, per capita income and economic growth. Applied Economics Letter, 3(9), 595–597.

43.

Ismail

N. W.

(2013). Innovation and high-tech trade in Asian countries. Paper presented at the International Conference on Recent Developments in Asian Trade Policy and Integration. https://www.nottingham.ac.uk/gep/documents/conferences/2013/malaysia-conference/normaz-wana-ismail.pdf

44.

Joseph

K. J.

(2013). Has trade been an engine of inclusive growth? India’s experience under globalization. Millennial Asia, 4(2), 135–157.

45.

Kayalvizhi

P. N.

, & Thenmozhi

(2018). Does quality of innovation, culture and governance drive FDI? Evidence from emerging markets. Emerging Markets Review, 34, 175–191.

46.

Keeble

, & Wilkinson

(2000). High-technology clusters, networking and collective learning in Europe. Routledge.

47.

Kim

(1980). Stages of development of industrial technology in a developing country: A model. Research Policy, 9, 254–277.

48.

Kroll

, & Liefner

(2008). Spin-off enterprises as a means of technology commercialisation in a transforming economy, evidence from three universities in China. Technovation, 28, 298–313.

49.

Kumar

, & Siddharthan

N. S.

(1997). Technology, market structure and institutions: Issues and policies for developing countries. United Nations University.

50.

Mani

(2000). Export of high technology products from developing countries: Is it real or a statistical artifact. In Mani

, & Romijn

(Eds.), Innovation, learning, and technological dynamism of developing countries (pp. 12–45). United Nations University Press.

51.

Martin

, & Bell

(2006). Technology spillovers from foreign direct investment (FDI): An exploration of the active role of MNC subsidiaries in the case of Argentina in the 1990s. Journal of Development Studies, 42, 678–697.

52.

Mason

, & Harrison

(2004). Does investing in technology-based firms involve higher risk? An explanatory study of the performance of technology and non-technology investments by business angels. Venture Capital, 6(4), 313–332.

53.

Mowery

, & Oxley

(1995). Inward technology transfer and competitiveness: The role of national innovation systems. Cambridge Journal of Economics, 19(1), 67–93.

54.

Pajooyan

, & Faghihnasiri

(2009). The effect of competitively on economic growth towards the endogenous growth pattern (in Persian). Economic Research of Iran, 13(28), 97–132.

55.

Patel

, & Pavitt

(1998). National systems of innovation under strain: The internalization of corporate R&D (Electronic Working Paper Series). Science Policy Research Unit.

56.

Pesaran

M. H.

, & Pesaran

(1997). Working with Microfit 4.0: Interactive econometric analysis. Oxford University Press.

57.

Pesaran

M. H.

, Shin

, & Smith

R. J.

(2001). Bound testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16(3), 289–326.

58.

Philips

P. C. B.

, & Hansen

B. E.

(1990). Statistical inference in instrumental variables regression with I(1) processes. The Review of Economic Studies, 57(1), 99–125.

59.

Porter

(1985). Competitive advantage, creating and sustaining superior performance. The Free Press.

60.

Porter

(1990). The competitive advantage of nations. Free Press.

61.

Prakash

J. V.

, & Nauriyal

D. K.

(2021). Integration and interdependence among equity markets in South Asia: Measuring through ARDL bounds approach. Millennial Asia, 12(2), 229–251.

62.

Romijn

, & Albaladejo

(2002). Determinants of innovation capability in small electronics and software firms in southeast England. Research Policy, 31, 1053–1067.

63.

Rong

, Liu

, Huang

, & Zhang

(2020). FDI, labor market flexibility and employment in China. China Economic Review, 61, 101449.

64.

Root

(1992). International trade and investment. South-Western.

65.

Roukanas

, & Karakostas

(2019). Is Japan a pioneer in high technology exports? The Romanian Economic Journal, 22(73), 2–18.

66.

Schneider

P. H.

(2005). International trade, economic growth and intellectual property rights: A panel data study of developed and developing countries. Journal of Development Economics, 78, 529–547.

67.

Schumpeter

J. A.

(1934). The theory of economic development. Harvard University Press.

68.

Scully

G. W.

, & Slottje

D. J.

(1991). Ranking economic liberty across countries. Public Choice, 69, 121–152.

69.

Seyoum

(2004). The role of factor conditions in high-technology exports: An empirical examination. Journal of High Technology Management Research, 15, 145–162.

70.

Shin

, Oh

, & Rho

(2019). Innovation and exports of Korean firms: How do they differ by size and industry? Millennial Asia, 10(2). 10.1177/0976399619879875

71.

Squalli

, & Wilson

(2011). A new measure of trade openness. The World Economy, 34(10), 1745–1770.

72.

Stoian

, & Iorgulescu

(2020). Fiscal policy and stock market efficiency: An ARDL bounds testing approach. Economic Modeling. 10.1177/0976399619879875

73.

Sultana

, & Turkina

(2020). Foreign direct investment, technological advancement, and absorptive capacity: A network analysis. International Business Review, 29, 101668.

74.

Tang

, & Zhang

K. H.

(2016). Absorptive capacity and benefits from FDI: Evidence form Chinese manufactured exports. International Review of Economics and Finance, 42, 423–429.

75.

Teixeira

A. A. C.

, & Fortuna

(2010). Human capital, R&D, trade, and long-run productivity. Testing the technological absorption hypothesis for the Portuguese economy, 1960–2001. Research Policy, 39, 335–350.

76.

Tingvall

, & Pohdahl

(2006). Is there really an inverted U-shaped relation between competition and R&D? Economics of Innovation and New Technology, 15(2), 101–118.

77.

Trischler

, Johnson

, & Kristensson

(2020). A service ecosystem perspective on the diffusion of sustainability-oriented user innovations. Journal of Business Research. 10.1016/j.jbusres.2020.01.011

78.

Vargo

S. L.

, Akaka

M. A.

, & Wieland

(2020). Rethinking the process of diffusion in innovation: A service-ecosystems and institutional perspective. Journal of Business Research. 10.1016/j.jbusres.2020.01.038

79.

(2000). Multinational enterprises, technology diffusion, and host country productivity growth. Journal of Development Economics, 62, 477–493.

80.

, & Wang

(1999). Capital goods trade and R&D spillovers in the OECD. Canadian Journal of Economics, 32(5), 1258–1274.

81.

Zhu

, & Jeon

(2007). International R&D spillovers: Trade, FDI, and information technology as spillover channels. Review of International Economics, 15(5), 955–976.

Augmented Dickey–Fuller (ADF) Test
Variables	Level		First Difference
Variables	Intercept	Trend and Intercept	Intercept	Trend and Intercept
HTRD	–2.605808* (8)	–1.786170 (8)	–4.311698*** (3)	–3.518148** (7)
RD	–0.470150 (0)	–2.670484 (0)	–10.49257*** (0)	–10.44232*** (0)
FDI	–2.146831 (8)	–2.485684 (8)	–6.048761*** (7)	–6.042710*** (7)
HC	–1.374551 (0)	–1.829747 (0)	–10.17736*** (0)	–10.15213*** (0)
RDHC	–1.343567 (0)	–1.406002 (0)	–10.01292*** (0)	–10.09381*** (0)
FDIHC	–3.499769*** (0)	–4.055738*** (0)	–10.02424*** (0)	–10.39800*** (0)
TO	–1.648714 (4)	–1.569296 (4)	–3.909650*** (3)	–3.927328** (3)
RDTO	–1.522028 (0)	–1.400890 (0)	–10.01116*** (0)	–10.18895*** (0)
FDITO	–3.931173*** (0)	–3.865524** (0)	–10.02007*** (0)	–10.17900*** (0)
EF	–1.524225 (0)	–1.273927 (0)	–10.00087*** (0)	–10.02648*** (0)
RDEF	–2.325117 (0)	–2.305724 (0)	–10.00642*** (0)	–9.978761*** (0)
FDIEF	–2.189357 (4)	–2.365999 (4)	–8.650271*** (3)	–8.615519*** (3)
INF	–2.449439 (0)	–2.405647 (0)	–10.00348*** (0)	–9.968524*** (0)
FD	–2.942100** (4)	–3.030531 (4)	–3.905307*** (3)	–4.061574*** (3)