Page 68 - 臺大管理論叢第33卷第1期

Value Creation and Capture in Developing Countries: The Driver and Mechanism of Offshore Outsourcing
Innovation

the industry where a focal firm operates also influences offshoring decisions (Lewin et
al., 2009). Finally, we use dummy variables to control the different types of innovation
activities outsourced offshore. We also include the calendar years from 1990 to 2009
when a focal firm conducted offshore outsourcing in order to control the possible temporal

effects during the observed period.

4. Results

4.1 Statistical Analysis and Hypothesis Test
Using the Heckman two-stage model with probit estimation, we test the relationships
between the hypothesized effects of variables and the location choices of outsourcing
innovation activities. The descriptive statistics of each variable and the bivariate

correlation between any two variables are listed in Table 2. As the analysis includes
an investigation of the firm’s location choice between strong and weak IPR protection
countries, the correlation results show a low potential of common method bias (Brannick,
Chan, Conway, Lance, and Spector, 2010). Nevertheless, we further conduct the CFA

analysis to detect the potential of common method bias (Brannick et al., 2010). The results
of the CFA analysis show a significant difference between the one-factor model and multi-
2
factor model (△χ = 1350, df = 15, p-value < 0.01), which indicates no common method
bias in this study. Table 3 includes the models predicting the firm’s choice between

each type of country for an outsourcing innovation activity. The estimated coefficients
represent the probability of choosing weak compared to strong IPR protection countries.
Statistically, a positive coefficient of a variable means an increase in the probability
of choosing the weak IPR protection countries while a negative coefficient means that

outsourcing an innovation activity is more likely to happen in the strong IPR protection
countries.
According to Model 1 (a null model containing only the control variables) of Table
3, the results show that there is the negative coefficient of turnover (β = -0.87, p-value

< 0.05). This indicates that the turnover of local employees increases the likelihood of
choosing countries with strong IPR protection to outsource an innovation activity. Cultural
similarity also increases the likelihood of outsourcing an innovation activity in the strong
IPR protection country (β = -1.96, p-value < 0.05). Specifically, the negative coefficient of

60