Page 69 - 臺大管理論叢第33卷第1期

NTU Management Review Vol. 33 No. 1 Apr. 2023

cultural similarity means that cultural difference between home country and host country
decreases the likelihood of choosing countries with weak IPR protection to outsource an
innovation activity. Further, a significant negative coefficient of the inverse Mills ratio
indicates that unobserved characteristics underlying a decision on offshoring influence the

location choices of those firms that did use the offshore outsourcing mode relative to those
that used the captive offshoring mode (β = -1.19, p-value < 0.05). In addition, the negative
effect suggests that a firm’s intention to use the offshore outsourcing mode based on

unobserved characteristics increases the likelihood of choosing countries with weak IPR
protection (Dolton and Makepeace, 1987).
According to Model 2 in Table 3, we find support for the hypotheses related to
the impact of available human capital on location choices for outsourcing innovation
activities. First of all, we include the variable of human capital “Low-cost talent” in

Model 2 of Table 3. In this model, the positive coefficient of low-cost talent means that the
availability of low-cost highly-talented human capital increases the likelihood of choosing
countries with weak IPR protection to outsource an innovation activity (β = 3.06, p-value

< 0.01). That said, if a country with weak IPR protection could provide abundant human
capital, it is more likely to be selected as an outsourcing location. Therefore, Hypothesis 1
is supported.
To test Hypothesis 2, we introduce one moderating variable (“Task specificity”) into
the estimation. We use the mean centered interactions between low-cost talent and task

specificity in estimating the probability of outsourcing an innovation activity to a country
with weak IPR protection. The results are listed in Model 3 of Table 3. In this model, the
interaction of low-cost talent with task specificity shows that the interaction between low-

cost talent and task specificity is statistically significant and positive (β = 2.59, p-value <
0.05). The positive estimated coefficient indicates that the design of task specificity can
enhance the effect of low-cost talent on the likelihood of choosing the country with weak
IPR protection for innovation outsourcing. Accordingly, Hypothesis 2 is supported.
To test Hypothesis 3, we introduce another moderating variable “Project modularity”

to estimate the probability of choosing a location to outsource an innovation activity. We
use the mean centered interaction terms between low-cost talent and project modularity.
The results are listed in Model 4 of Table 3. The results show that the interaction between

low-cost talent and project modularity is statistically significant and positive (β = 2.57,

61