You should surely have a look at this site of MyC4, which is probably the most innovative Micro-finance project for financing investment in Africa. It brings private savers and debtors directly together.
Even though there is no guarantee of repayment nor an enforcement mechanism, loss of reputation (and accordingly bad rating for a potential new application for funds) and an incentive scheme to pay early (with lower interest) have so far allowed a pretty good return according to MyC4.
Up to now, MyC4 carried the exchange rate risk but it plans to move away from this model.
My bet is that once this happens creditors will tend to lend money to countries with fixed exchange rate regimes, since the little private saver has no time to deal with predicting exchange rates and is more likely to believe that for the time of his/her investment there will be no devaluation.
As a consequence interest rates for loans in those countries are likely to be lower. So much about the (ir-) relevance of the nominal exchange rate regime.
It would be interesting to analyze the dataset in order to see how people react to the introduction of exchange rate risk. It could be used for instance to infer how many percentage points in terms of expected return creditors are willing to sacrifice in order to not be exposed to exchange rate risk.
MyC4 will possibly explore this dimension to provide insurance against exchange rate variation. This would allow again the equalization of the expected returns.
Personally, I would prefer to have the option to - rather than pick a single project - be allowed to spread my little savings that I have over all the projects and be willing to accept a below average return on this portfolio...Who knows, maybe MyC4 will provide this (investment fund style) option at a later stage.