Robustness test. The robustness of the product must also be determined in the course of the framework of testing, in order to detect manufacturing variations. In this context, it can be necessary to maintain actual characteristics of some prototypes deliberately outside the nominal values, and even in the threshold range or outside the limits of tolerance. 8.3
Robustness test. In this section, we conduct robustness tests on the impact of political and economic factors on SOE-involved alliances. To capture the overall impact of those country-level factors, we include all of them in the regression model. Since political and economic factors could be correlated, we construct a single index by using the principal component analysis of these seven key variables (“First Principal Component”) suggested by Xxxxxxx and Xxxx (2017). The details of the approach are shown in Online Appendix Table B3. We take the first principal component to build the index since the eigenvalue is greater than 1 and accounts for 26.6% of the variation across the seven variables. The significantly negative coeffieint of the first principal component shown in the model suggests the robustness of our results. Besides, we also consider that the impact of political and economic factors on SOE- involved alliances might be driven by the large proportion of Chinese SOE-involved deals compared with other countries.25 To mitigate this concern, we remove the alliance deals with Chinese local partners from the analysis. We find that our main results are unaffected. The robust results are reported in Online Appendix Table B4. Furthermore, we control for the impact of bilateral investment treaties (BITs) between two signatory countries on the cross-border activities, as BITs could protect and stimulate the foreign investment such as cross-border mergers (Xxxxxxx et al., 2021) and international collaboration in producing innovation (Bian et al., 2021). We add a dummy variable (“BIT Dummy”) in our baseline model, which equals one if the BIT is in place in a given year between the paired countries, and zero otherwise.26 The results in Online Appendix Table B5 show that impact of political and economic factors on SOE-involved alliances remains significant after controlling for the BIT effect. Also, the significantly positive coefficient of BIT Dummy suggests that the signature of BIT could encourage more foreign firms to ally with local SOEs 25 Table 1 indicates that the Chinese SOE-involved cross-border alliances account for 27.23% of the total number of SOE-involved deals in the world. 26 If BIT are terminated between the paired countries during the sample period, the dummy is equal to zero for the period before the signature date and for the period after the termination date. than non-SOEs in the host countries.
Robustness test
