“…Second, for endogeneity concerns related to variables that are potentially missing from our model, we conducted the Hausman specification test (Davidson and MacKinnon, 1992) using instrumental variables (IV) as correlated with the focal predictor, but not the error term (Angrist et al, 1993). Following Hunt et al (2021), a set of three instruments were used: (i) the difference in the home market and international market population growth rates among children aged five to fifteen, which is a forward-looking indicator of both labour availability and the future demand for commodities; (ii) the firm founding rate differences between each firm’s home market and aspirant international market (GEM, 2018), which provide a proxy for formal and informal institutional encouragement for small businesses (Veciana and Urbano, 2008); and (iii) the survey responses of owner-operators in the study concerning succession planning for their respective businesses, which proxies the forward-looking nature of transactional decision-making (Hunt et al, 2021). The three instruments were found to be jointly significant predictors of internationalisation, ( F 4, 14,348 = 18.47, p ≤ 0.01), referrals ( F 4,14,348 = 12.90, p ≤ 0.01) and operational performance ( F 4, 14,348 = 9.24, p ≤ 0.05), thereby establishing satisfactory IV strength.…”