In small and medium-sized enterprises (SMEs), which suffer from lack of resources to search for and to analyze information, internationalization decisions are often concentrated in a single decision maker – the export manager, who often has limited access to information. This study investigates whether the personal characteristics of the decision maker – in particular his/her age, human capital (educational background, foreign language expertise and international experience) and his/her social capital (personal and professional relationships that facilitate his/her access to information sources and other resources) – affect the export decisions of SMEs, specifically regarding earliness of export activity, growth rate of international sales, number of countries served and intensity of international sales. This study examines both the direct effects of the manager s characteristics and the mediated effects through risk propensity and risk perception (about exports). The literature presents arguments and empirical results about the impact of the manager s characteristics on the path of internationalization of SMEs that are partially contradictory. Data were collected through an online survey sent to more than 6,000 exporters, whose e-mails were obtained from the Brazil4export (http://www.brazil4export.com/) database, provided by CNI (National Confederation of Industry). A total of 241 questionnaires were responded, of which 118 were discarded (because they corresponded to companies (large) or respondents (not export managers) not eligible for the study or the questionnaire had more than 15 per cent missing data) and there remained 123 cases; of these, only 71 corresponded to companies whose export manager who answered the questionnaire had been working at the company for at least three years when the company started exporting. The model was estimated with the PLS (Partial Least Squares) technique, which is robust to small samples; three control variables were used - industry, company size and degree of satisfaction with the performance of domestic operations. The results of the study suggest that age have negative impact (as expected) and statistically significant about the speed of globalization, but not significant on the propensity to risk and the degree of intrancionalização. All the components of human capital has positive impact (as expected), but statistically not significant on the propensity to risk, the degree of internationalization and the speed of internationalization. With respect to the components of the social capital, only relationships (but not information sources) showed statistically significant effect on the share capital and this showed significant positive effect on the propensity to risk (as expected) but not significant on the perception of risk or about the speed of internationalization and the degree of internationalization. Propensity to risk and risk perception were not significantly associated with internationalization speed and also with the degree of internationalization. The small sample size implies a reduced statistical power, that is, the probability of rejecting the null hypothesis if it is in fact false. The G (asterisk) Power software indicated that for the explanatory model of this study and this sample size (71 cases), the power of the test was only 71 per cent to detect medium effects (f2 large, equal 0.15) at the 10 per cent significance level.