While economies of scope of lending and mobilizing deposits in banking are justified theoretically (Diamond, 1984) and found empirically (see Saunders, 1999), in microfinance, the existence and the magnitudes of scope economies has not been investigated. We use a semi parametric smooth coefficient model to estimate these economies using the largest publicly available dataset of over 2,800 annual observations from MFIs across the world. We accommodate the outreach objective of Microfinance Institutions (MFIs) by measuring outputs as the number of clients (borrowers and savers) rather than the traditional volume of loans and deposits and compare the results from the two cost functions. We find substantial scope economies of 47 percent on average. The results also indicate that these economies are entirely due to shared fixed costs since MFIs could not profitably use information they collect from savers to improve their lending products and vice versa. In fact, we find large negative variable cost complementarities. Overall scope economies are the largest in rural banks and coops, while lowest in banks. We further find that, while very few of MFIs in Eastern Europe and Central Asia offer savings, the potential scope economies that they can realize are by far the largest. Finally we compare the two methods and find that while on the average the traditional cost function approach would underestimate scope economies, it would overestimate them for MFIs in Africa and all parts of Asia, as well as for NGOs and non-bank financial institutions.