thanks for your quick reply.
No doubt that the package of training and finance works better, but I wonder from where you take the certainty that training alone does not work. I would rather conclude that the results are very mixed. McKenzie and Woodruff 2012 find in their literature review very little evidence, but Mano et al 2011 find some very positive impact on business performance in their experiment testing SIYB in Ghana (unfortunately they did not go for labour market outcomes). The meta regression analysis of Cho/Honorati 2013 assessing some 35 impact evaluations on entrepreneurship development also highlights that training alone can work. Am I too biased?
With regard to charging a fee I think your argument of risking to exclude the ones who most need the training goes too short. Charging a fee is considered a best practice among practitioners as it helps to increase financial sustainability of local training providers and as it is meant to serve as a filter to select committed participants. But there is no solid evidence. Therefore, I think that future experiments should include a component testing and quantifying the willingness to pay in order to answer the question whether charging a price affects selection and impact.