Speaker
Description
This paper analyses the impact of mobile money adoption on informal firms performance in 8 African countries. Firm performance is approximated by firms’ total revenues,
total revenues per employee and number of employees. Measuring the impact of mobile
money adoption raises problems of endogeneity and selection bias, since the choice of
adopting mobile money (or not) is not random. Following recent developments in the
literature concerning these issues, we apply the entropy balancing matching method to
transform our data so that the treatment group (firms that have adopted mobile money)
and the control group (firms that have not adopted mobile money) become similar in
terms of moments of our variables of interest. We then measure the impact of mobile
money adoption using weighted least squares with weights generated by balancing
entropy. Finally, we validate the sensitivity of our results by performing the Oster (2019)
test for variable omission bias. Interpreting our most complete regressions, we can
say that mobile money adoption increases annual sales for our sample of informal
firms by 41% and annual sales per employee by 31%. We also check whether selection
on unobservables could affect our results by doing Oster (2019) test. We find that
(unobservables) biais adjust treatment effect has the same magnitude as our complete
regressions and unobservables selection is not sufficient to drive a zero effect of mobile
money on firms’ performance
