**3. Results**

Both Fixed- and Random-Effects models are used to obtain results for both companysize cohorts. The results for both the FE and RE analysis are listed in the Appendix A (Tables A1 and A2). To find out whether the RE model provides new information in addition to the results from the FE estimation, the Hausman test, which tests the presence of individual effects by comparing the FE and RE models' coefficients, was performed. If there are no significant differences, the individual effects are random and thus either of the estimators can be used [40]. The alternate hypothesis (*p*-value < 0.05) is that the FE and RE coefficients are different from each other and in such a case, only the FE estimator is consistent.

Tests on heteroskedasticity and autocorrelation indicate that they were present regularly. Arellano's (1987) [41] and White's heteroskedasticity robust standard errors [42] are used in the analysis when heteroskedasticity is present (see Tables A1 and A2).
