Who should receive a cash transfer?, or the paradox of micro-credit

Let’s say you send regular money to a poorer individual in another country.  You might wonder what are the possible rates of return on those funds, and furthermore does that analysis shape to whom you should give the money?

I don’t quite believe the argument I am about to write out, but I can’t yet find the flaw in it either.

Let’s say you find an individual borrowing micro-credit.  It is well-known that rates of interest on these loans often run between 50 to 100 percent, annualized, and furthermore many individuals/families dip into these markets frequently.  Furthermore, very high quality RCTs by Duflo et.al. and Dean Karlan show that micro-credit is not on average harming the families who borrow.

That implies these economies — at least in some their corners — have investments and/or liquidity deployments worth at least fifty percent per annum.  For simplicity, I will use an estimate of fifty percent.

Go to a borrowing individual and give him/her some money for free.  If micro-credit is no longer necessary, you have given that individual a high return.  If it was a cash-free loan, the return to that person would be fifty percent.  By simply giving the money away, it would seem the rate of return would be at least 2x that, or at least one hundred percent.  That is pretty good too.  Of course that individual might stay in the micro-credit market (post-gift), but that implies there are still worthwhile additional uses for the marginal liquidity.  And we’ve already seen that micro-credit does not usually harm those who use it.

So you’ve generated returns of one hundred percent or more with your cash transfer.  This does not require heroic acts of entrepreneurship, merely that the individual was previously a responsible user of micro-credit.

It also requires that these individuals be reasonably conscientious, and do not simply squander their new-found wealth.

But the core recipe is to give to conscientious current borrowers, for very high rates of return.

What is wrong with this argument?