Page 80 - 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

providers (Baldwin and Henkel, 2015; Brusoni and Prencipe, 2001; Fixson and Park,
2008). This is another type of task design with interdependencies among different tasks.
By introducing these two mechanisms used to manage offshore outsourcing innovation, we
shed some light on the recent research into global outsourcing and offshoring strategies.

5.3 Managerial and Policy Implications
Our findings have several managerial implications. Our estimates of the offshore

outsourcing locations indicate that within our sample, sourcing locations are influenced
not only by cost advantages but also by human capital in host countries. This suggests that
developing countries can be selected as the sourcing locations when they are able to supply
low-cost and high-skilled talent. In addition, our finding about the positive impact of task
specificity and the positive effect of project modularity on the relationship between human

capital and location choice also make important reading for managers contemplating
offshore outsourcing innovation, especially for firms that plan to outsource innovation in
developing countries with weak IPR protection. Our findings should also be of interest to

policy makers. Our study suggests offshore innovation outsourcing in developing countries
may not only be driven by cost advantages but also by a firm’s desire to obtain human
capital. Thus, programs that would enhance human capital in a firm’s home country might
be a way of reducing its propensity to outsource innovation activities abroad. This can be
particularly applicable for the knowledge-intensive industries.

5.4 Limitations and Future Directions
In addition to our contributions and implications to research on offshore innovation

outsourcing activities and location choices, we need to acknowledge several inherent
limitations which might affect the generalizability and validity of the empirical findings
in this study. First, we use the ORN survey data to measure major variables such as low-
cost talent, turnover, task specificity, and project modularity. The valid measures might be
limited in this study. In addition, some potential errors of measurement in the survey data

might also influence the estimated effects of the major covariates. Future research could
incorporate other variables to measure latent concepts and include more control variables
to improve these limitations and overcome bias. For example, future researchers may

build a more comprehensive model by also considering the effect of informal institutions

72