Reworking the final chapter, I realized I needed to understand better what is called the "financing structure" of microfinance institutions (MFIs)---where they get the money to lend to clients. In general, MFIs can borrow from big banks and investors or issue bonds; take deposits (savings) from clients; and accept equity investments, which are ownership stakes that earn a share of the profits. A theme in the chapter is the danger that the first category, wholesale borrowings, is crowding out the second, savings, resulting in MFIs that underemphasize microsavings and overemphasize microcredit.
(Technical note: "Equity" is a residual category. If equity investors put in $100 million, their equity could change to $110 million or $90 million a month later depending on the gains and losses of the MFI. So $100 million in equity is not $100 million received from investors.)
In countries where microfinance is oldest, the financing structure tends to be diverse. Where microcredit is new and growing fast, borrowings dominate. The latter includes Nicaragua, Pakistan, Bosnia, and Morocco, the four countries where microcredit bubbles popped last year. It also includes India, where hot growth has been financed primarily by borrowings from domestic banks, which are in turn propelled by the Indian government's "priority sector" lending requirements. I ended up expanding on this point in a revised chapter 8 (.docx and .pdf).
Here's a graph I added to chapter 8 (with help from Sparklines for Excel). I took the top 25 microfinance countries by number of outstanding loans, then sorted them by the percentage of finance that comes from borrowings: