This study estimates how many continuing MFI clients in India have crossed the USD 1.25 PPP (per capita) a day consumption threshold from below. The time period considered for our study is 1990-2010. We find that, at an all India level, (approximately) a net of 12% ( (9 million) of all MFI clients have crossed the USD 1.25 a day consumption threshold from below. To arrive at the all India level estimate, we first divide the country into 6 distinct regions (North, South, East, West, Central and North East). We then identified the top 14 states whose MFI clientele is at least 1% or more of the all India total. These 14 states together have approximately 95% of the total MFI clientele in India. Thus the states are fairly representative of all the MFI clientele in India. MFIs in India are not homogenous. Significant differences exist across MFIs either because they have different organizational structure or they practise different lending models. Therefore, the MFIs are classified into 9 cells. Each cell is a combination of a loan type (joint liability, individual liability or self help group) with any one of the three organizational types (NGO, NBFC, others). For each of the states, MFIs were identified belonging to each cell. There were 18 cells considered for each State-9 in rural and 9 in urban. There were 27 MFIs and 6 SHG Bank linked NGOs who participated in the study. The two most populous cells are NGO-SHG and NBFC-JLG combination. Households were selected randomly from the clients list of the MFIs and surveyed. We surveyed 15,205 households. However, after omitting the missing observations, we were left with 14,746 households in all. The household questionnaire captured the current asset distribution as well as the asset distribution at the time of their joining the specific MFI. The latter was found out using the recall method. The estimates were obtained by scoring the population from the collected data. Different scorecards for rural and urban households were developed for each of the 14 States. These scores are calibrated to the person’s likelihood of being above or below a given consumption threshold. Once the estimates for each cell were obtained, the population estimates were arrived at by using the population for each cell as the multipliers. We used the Statistical package Strata 8.E for the entire analysis.