I've read several - but not all - of those blogs.
If I may be so bold, the criticisms seem to revolve around these facts:
1) Many loans are disbursed to lenders before the Kiva lender chooses to make a loan. So you're actually reimbursing the microcredit organisation for a loan already made, rather than causing a loan to be initiated. (This is obvious if you look at the loan details; the "date funds disbursed" is often already past.)
2) Like child sponsorship, the Kiva model of microcredit relies on creating a connection between the donor and a particular individual, which is expensive and largely artificial (ie the borrower won't know that you are the one who's made the loan to them).
3) The loan repayment information is assumed, ie Kiva assumes that the borrower is repaying loans on schedule unless informed otherwise.
To me, all of these reflect that there's a trade-off necessary between creating the real nexus between borrower and lender, and being efficient (ie minimise expenses incurred in generating that connection).
I think that Kiva has the balance about right, and is pragmatic. The people that you "lend" to really did receive a microcredit loan for the terms specified. The only expense involved in connecting a lender to a borrower is that when loans officers take loan applications, they also take a photo and some details of the borrower, for posting on Kiva to get back the funds that they'll (usually) have already loaned to the borrower.
Delaying disbursement of funds until Kiva lenders have filled the loan, and requiring updates from each borrower to be posted on Kiva every time there's a repayment due, would both incur expense and reduce efficiency. (Does it make sense to incur a $30 expense to find out whether an $80 payment has been made on time?)
I think Kiva provides "enough" of a connection between borrower and lender for it to appeal to people and be successful in soliciting funds, but not so much that the costs of the connection become substantial relative to the size of the loan.
I think there are criticisms of Kiva, but many of them are from people who are even more idealistic than I am.
I totally agree that the ideal model would be for people to just give freely, and say "spend it where it's most effective", without the need for feedback as to precisely how the money was spent. That's the most efficient model. But in the real world, that model isn't as successful; people simply don't give.
But thanks for opening up the debate, Jas. I still think Kiva's infinitely preferable to child sponsorship, which has the balance far too strongly swayed towards the "connection" rather than the "purpose of the funds" end of things. If Kiva is "exploiting" people to be the face of third-world poverty, they are at least adults who've freely chosen to participate, rather than children, and the funds are a loan (dignity-preserving) rather than a hand-out.
