Over the past 30 years, microfinance (MF) has grown to a $30 billion industry involving over 3,000 organizations. Half of these microfinance institutions, however, have nothing more than anecdotal stories about the impact of MF on poverty. The connection between MF and poverty eradication is far from resolved. This paper addresses an ongoing debate by drawing upon two bodies of literature: poverty eradication and microfinance measurement. It articulates a causal model of the effects of MF on poverty that is unique in three regards. First, it measures the effect on socioeconomic status (SES) using for the first time a reflective multi-dimensional construct in contrast to deploying a financial surrogate for SES such as purchasing power parity (PPP). Second, it applies variables in measuring SES that are viewed as reflections of their origin – the level of poverty. This draws upon factor analytic techniques in contrast to earlier stepwise approaches. Third, on the cause side, in contrast to positing a direct effect the model evaluates MF for its moderating effects on the causal positive association between income and SES. The empirical setting for validating the proposed model is Malawi, a country of extreme poverty in central Africa. We transform an existing World Bank data set on household income and poverty in Malawi into a causal model using structural equation modeling (SEM). Our analysis shows that MF has a significant interaction with income in its impact on SES, thus demonstrating its instrumental value in reducing poverty. Contrary to earlier studies we find no differences across gender or income levels for the impacts of MF. Also contrary to previous literature, the effect is different between urban and rural households. The proposed approach offers a handy protocol to evaluate rigorously the impacts of MF or other poverty reducing efforts in other countries and contexts enabling cumulative empirical findings. On a practical level the study offers some guidelines how to target future microfinance operations.