Are seminars on export promotion effective? Evidence from a randomised controlled trial

Yu Ri Kim, Yasuyuki Todo, Daichi Shimamoto, Petr Matous

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)


This paper investigates the impacts of informational and motivational seminars on export promotion targeting small and medium enterprises in the traditional apparel and textile clusters in Vietnam. To control for biases due to self-selection, we conducted a randomised controlled trial and invited randomly selected firms to participate in 1-day seminars. Because only some of the invited firms participated in the seminars, we employ an instrumental variable approach in which the dummy variable for a random invitation is used as an instrument for quantifying the effects of participation. We find that seminar participants were more likely to sense the difficulties of the export procedures and were, on average, unlikely to start exporting in the short run. However, the seminars encouraged large firms, which possibly embody higher productivity and absorptive capacity, to start exporting shortly after the seminars. This effect was not sustained in the long term without additional stimuli. We have also identified spillover effects of participants on non-participants through informal information networks within the cluster. Overall, the results suggest that the provision of information about export is useful only to adequately productive larger firms with a sufficient absorptive capacity.

Original languageEnglish
Pages (from-to)2954-2982
Number of pages29
JournalWorld Economy
Issue number11
Publication statusPublished - 2018 Nov


  • Vietnam
  • export promotion
  • impact evaluation
  • randomised controlled trial
  • small and medium enterprises

ASJC Scopus subject areas

  • Accounting
  • Finance
  • Economics and Econometrics
  • Political Science and International Relations


Dive into the research topics of 'Are seminars on export promotion effective? Evidence from a randomised controlled trial'. Together they form a unique fingerprint.

Cite this