Extension of this study

Based on the literature review, this study adopts and extends the theoretical framework and the research method of prior research as follows:

2.5.1 Adoption from prior research (1) Theoretical and research framework

To keep the theoretical model tractable, the simplified assumptions of the models introduced by D’Aspremont and Jacquemin (1988) and Kamien et al. (1992) are maintained. The main assumption is that the firms play a two-stage game. In the R&D stage, the firms simultaneously decide on their level of R&D investments. In the output stage, the firms in each industry decide on the quantity they will produce and sell on the market. By backward induction, I derive the subgame perfect Nash equilibrium.

Based on the two vertical industry models, Atallah (2002) includes four R&D scenarios: R&D competition, vertical R&D cooperation, horizontal R&D cooperation, and generalized R&D cooperation. In this study I adopt his R&D cooperation scenarios in my theoretical model.

z I also follow the approach proposed by Steurs (1995) to simulate and compare the equilibrium R&D investments, R&D outputs, and profits for different cooperative scenarios.

(2) Research method

z Hagedoorn and Schakenraad (1994) pioneer in R&D cooperation empirical research by using a systematic collective database. In this study I follow their method to collect R&D cooperation data in Taiwan’s high technology industries.

z To capture the frequency and intensity of R&D cooperation, I follow Stuart’s (2000) approach and use the number of R&D cooperation to measure R&D cooperation intensity.

z Prior studies only measure uncertainty as the dispersion from the mean and do not detect the ordering of the data points. However, it is unable to detect

approach to solve this problem..

2.5.2 Complement of prior research (1) Theoretical and research framework

z Incremental to prior literature (e.g. D’Aspremont and Jacquemin 1988;

Kamien et al. 1992; Inkmann 2000; Atallah 2002; Ishii 2004), this paper introduces two-industry, n-firm-per-industry models to facilitate generalization of analytical results.

(2) Research method

z Few empirical papers comprehensively examine the direct and indirect relationships among R&D cooperation, R&D investments, R&D outputs, and financial performance, and most of them use survey data to test the theoretical hypothesis (e.g. Peters and Becker 1997-98; Bayona et al. 2001;

Cassiman and Veugelers 2002; Becker and Dietz 2004; Belderbos et al.

2004). In this study I develop an integrated R&D cooperation—innovation—financial performance framework and apply path analysis to analyze the direct and indirect relationships among R&D cooperation, R&D investments, R&D outputs, and financial performance.

z This paper first includes four types of R&D cooperation, including horizontal R&D cooperation, vertical R&D cooperation, generalized R&D cooperation, and R&D competition, to empirically test how they influence companies’ R&D investments, R&D outputs, and financial performance in Taiwan high-technology industry. I also examine how the intensity of different types of R&D cooperation is influenced by three factors:

knowledge spillovers, uncertainty, and absorptive capacity.

z To avoid sample selection bias, I apply the Heckman two-stage model and treatment effects model to test the impact of R&D cooperation on R&D investments, R&D outputs, and financial performance. In addition, to incorporate multiple level variables into an empirical test, I use the HLM (Hierarchical linear model) to examine the determinants of R&D cooperation.

z Taiwan is a core innovator internationally, according to the World Economic

R&D cooperation and firm performance in Taiwan with small databases (e.g.

Chang 2003; Sher and Yang 2005). This study is the first to use Taiwan’s high-technology industries as a research sample and to thoroughly explore the relationships among R&D cooperation, R&D investments, R&D outputs, and financial performance in Taiwan.

z In this study I argue that higher frequency of inter-industry or intra-industry strategy alliance implies higher knowledge spillovers among firms.

Therefore, I first use the number of strategy alliance for each industry to proxy knowledge spillovers. In addition, a common feature in the prior R&D cooperation literature is the absence of uncertainty. Therefore, I will discuss the relationship between uncertainty and R&D cooperation in this paper.

See Table 8 for the summary of the extension of this study.

Table 8: Summary of the extension of this study

Item Adoption from prior research Complement of prior research

Theoretical and research framework

1. To keep the theoretical model tractable, the simplified assumptions of the models introduced by D’Aspremont and Jacquemin (1988) and Kamien et al. (1992) are maintained.

2. In this study I adopt Atallah’s (2002) four R&D cooperation scenarios, including R&D competition, vertical R&D cooperation, horizontal R&D cooperation, and generalized R&D cooperation, in my theoretical model.

3. I follow the approach proposed by Steurs (1995) to simulate and compare the equilibrium R&D investments, R&D outputs, and profits for different cooperative scenarios.

This paper introduces two-industry (upstream and downstream industries), n-firm-per-industry models to facilitate

generalization of analytical results.

Research method

1. In this study I follow Hagedoorn and Schakenraad’s (1994) method to collect R&D cooperation data in Taiwan’s high technology industries.

1. In this study I develop an integrated R&D

cooperation—innovation—financial performance framework and apply path analysis to analyze the direct

Item Adoption from prior research Complement of prior research 2. I follow Stuart’s (2000) approach and use the number of

R&D cooperation to measure R&D cooperation intensity.

3. I apply Dess and Beard’s (1984) approach to measure uncertainty which considers variation from a time trend.

and indirect relationships between R&D cooperation, R&D investments, R&D outputs, and financial performance.

2. This paper first includes four types of R&D cooperation, including horizontal R&D cooperation, vertical R&D cooperation, generalized R&D cooperation, and R&D competition, to empirically test how they influence

companies’ R&D investments, R&D outputs, and financial performance in Taiwan high-technology industry. I also examine how the intensity of different types of R&D cooperation is influenced by three factors: knowledge spillovers, uncertainty, and absorptive capacity.

3. To avoid sample selection bias, I apply the Heckman two-stage model and treatment effects model to test the impact of R&D cooperation on R&D investments, R&D outputs, and financial performance. In addition, to incorporate multiple level variables into empirical test, I use the HLM (Hierarchical linear model) to examine the determinants of R&D cooperation.

Item Adoption from prior research Complement of prior research

4. This study is the first to use Taiwan’s high-technology industries as a research sample and thoroughly explores the relationships between R&D cooperation and firm

performance in Taiwan.

5. In this study I first use the number of strategy alliances for each industry to proxy knowledge spillovers. In addition, I discuss the relationship between uncertainty and R&D cooperation.

Chapter 3: Theoretical model and hypotheses development