‘Learning to export’ and ‘learning to innovate’: Revisiting the relationship between innovation and exports in African ï¬ rms

This paper examines the relationship between innovation and export performance for African i¬ rms. We use Tobit simultaneous equation full information maximum likelihood (FIML) model with selection on a crosssectional dataset from the World Bank’s Enterprise Surveys for 28 African economies. The paper provides new evidence of a two-way positive relationship between innovation and export performance in African i¬ rms: innovation is important for both the ability to export (export propensity) and for export intensity, while exporting also increases the likelihood of innovating. These ei¬€ects are driven mainly by direct exports and apply to both product and process innovation. We argue that these results point to a two-way relationship in which innovation enables i¬ rms to ‘learn to export’, while i¬ rms also ‘learn to innovate’ through exporting. A higher share of foreign ownership in i¬ rms, as well as i¬ rms having an internationally recognised quality certii¬ cation strengthen the positive ei¬€ects in both directions.